mirror of
https://github.com/BerriAI/litellm.git
synced 2026-09-15 23:31:29 +00:00
432 commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
yucheng-berri
|
61d32c9aac
|
fix: handle explicit outputInfo: null in Vertex AI batch response (#34473)
* fix: handle explicit outputInfo: null in Vertex AI batch response
Vertex AI can return HTTP 200 for a create_batch/get_batch call with an
explicit "outputInfo": null body (the output directory is assigned
asynchronously and may not be populated yet at response time).
_get_output_file_id_from_vertex_ai_batch_response did:
response.get("outputInfo", OutputInfo()).get("gcsOutputDirectory", "")
dict.get(key, default) only substitutes default when the key is absent,
not when it is present but explicitly None, so this crashed with:
AttributeError: 'NoneType' object has no attribute 'get'
surfaced to callers as an opaque openai.InternalServerError 500 from
litellm.create_batch()/retrieve_batch() for any Vertex AI batch job,
regardless of whether the job ultimately succeeds.
Fixed by guarding with `response.get("outputInfo") or OutputInfo()`,
matching the existing null-safe pattern already used by the sibling
_get_input_file_id_from_vertex_ai_batch_response for inputConfig. The
existing outputConfig fallback branch (a few lines below) already
handles this case correctly once it's reachable - it just never was.
Added 2 regression tests covering outputInfo: null with and without an
outputConfig fallback available.
* test: drop explanatory comment from regression test
---------
Co-authored-by: htourinho-clgx <htourinho@cotality.com>
|
||
|
Yassin Kortam
|
f6a1050cbf
|
fix(vertex): incrementally parse accumulated Gemini stream JSON to prevent multi-value wedge (#34320)
The accumulated-JSON fallback ran json.loads over the whole buffer after every fragment and, on failure, kept the buffer without resetting it. A buffer that ever held more than one concatenated JSON value could never parse (json raises on trailing data), so it returned None on every subsequent chunk while growing without bound - an unrecoverable per-request CPU spin. Parse one value at a time from the front with raw_decode and keep the remainder, draining trailing values on later calls and at end of stream. |
||
|
Yassin Kortam
|
692b22655e
|
fix(logging): stop scheduling sync failure_handler concurrently with async_failure_handler (#34306)
The async streaming error paths fired the sync failure_handler in a thread and the async_failure_handler via create_task at the same time, so both mutated the shared logging object concurrently and could crash pydantic-core. Route failure logging through a single guarded dispatch_failure_handlers, so the sync handler only runs after the async one completes. |
||
|
Mateo Wang
|
9ab3847c0b
|
fix(tests): remove importlib.reload of http_handler that breaks client injection in later tests (#34336)
The huggingface embedding test fixture reloaded
litellm.llms.custom_httpx.http_handler, creating a new HTTPHandler class
object. llm_http_handler keeps the class captured at import time, so any
test running later in the same process that injects a client built from
the reloaded class fails the isinstance check and the mock is silently
discarded, causing a real network call. Under pytest-xdist loadscope this
surfaced as a deterministic failure of
test_accept_header_in_completion_request_jwt whenever an unrelated PR
shifted worker distribution.
Also removes the same reload pattern from the vertex rerank integration
test (both were previously removed in
|
||
|
mateo-berri
|
fa025fc474 | chore(tests): replace a customer name and domain with neutral placeholders | ||
|
Mateo Wang
|
1e741094fe
|
Merge pull request #33807 from BerriAI/litellm_vertex_azure_midsys
fix(vertex,azure): model-aware mid-conversation system for Claude /v1/messages |
||
|
mateo-berri
|
8b1a19fb02 | test: give cost-map guard next() a default so a renamed rule fails with a clear assertion | ||
|
mateo-berri
|
23b5b7d199 |
fix(vertex,azure): model-aware mid-conversation system for Claude /v1/messages
Azure AI Foundry and Vertex AI serve Claude on the first-party Anthropic
Messages contract, which was verified live to be byte-identical to
api.anthropic.com: a leading role:"system" entry in messages is rejected on
every model ("messages.0: use the top-level 'system' parameter"), and a
mid-conversation role:"system" reminder is accepted in place on Claude 4.8+/5
but 400s on Claude 4.7 and older ("role 'system' is not supported on this
model"). This is the same contract Bedrock Invoke already handles model-aware
(PRs #32578/#32831/#32882); Vertex and Azure did no hoisting at all, so a Claude
Code session on an older Vertex/Azure Claude model hard-400s on its reminder
turns, and the only thing sparing 4.8+/5 was that nothing was hoisted
Extract Bedrock's model-gated normalization into the shared
AnthropicMessagesConfig base as _normalize_system_role_messages and call it from
the Vertex and Azure messages configs. Flagged models (4.8+/5) hoist only the
leading run of system entries and keep mid-conversation reminders in place so
the top-level system prefix stays byte-identical and the prompt cache is
preserved; unflagged models hoist every system entry so the request returns a
completion instead of a 400
Add supports_mid_conversation_system to the azure_ai and vertex_ai Claude 4.8+/5
cost-map entries. Exact cost-map hits win over the claude-mid-conversation-system
fallback rule, so without the explicit flag those models would be treated as
unsupported and hoist every reminder, collapsing the prompt cache (the exact
customer regression). A per-provider test guards this so future 4.8+/5 entries
cannot silently miss the flag
Closes the Vertex/Azure gap from the customer RCA
|
||
|
devin-ai-integration[bot]
|
07e07e6e2b
|
fix(vertex_ai): exclude Gemini Google Search grounding tokens from input token billing (#33742)
* fix(vertex_ai): exclude Google Search grounding tokens from Gemini input token billing Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> * test(proxy): stub get_configured_token_limits on mocked routers Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> --------- Co-authored-by: Krrish Dholakia <krrishdholakia@berri.ai> Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
devin-ai-integration[bot]
|
4cfc987f56
|
fix(vertex_ai): surface Gemini grounding toolUsePromptTokenCount in Usage (#33533)
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> |
||
|
mateo-berri
|
ce3bf2d839 | fix(gemini): map video response modality instead of MODALITY_UNSPECIFIED | ||
|
mateo-berri
|
abc38935fa | fix(anthropic): override custom_llm_provider in provider config subclasses so capability probes use the right namespace | ||
|
Mateo Wang
|
46d9742950
|
fix(vertex_ai): return create_vertex_url result directly for openai-path partner models with custom api_base (#32380) | ||
|
Mateo Wang
|
b8248a21d2
|
fix(vertex_ai): build full request path when custom api_base has no path (#32367)
Some checks failed
CodSpeed Benchmarks / benchmarks (push) Waiting to run
GitHub Actions Security Analysis / zizmor (push) Waiting to run
Terraform Provider / gofmt, vet, build, test (push) Has been cancelled
Terraform Provider / Provider endpoints vs proxy OpenAPI schema (push) Has been cancelled
|
||
|
Mateo Wang
|
43b0a25f07
|
feat(vertex_ai): add Google Cloud Speech-to-Text Chirp 3 transcription support (#32274)
* fix(llm_http_handler): send dict transcription request data as a JSON body
httpx form-encodes dicts passed via data= and silently ignores json=, so the
generic audio transcription path never actually sent a JSON body. No provider
hit this before; JSON-body speech APIs need it.
* feat(vertex_ai): add Google Cloud Speech-to-Text Chirp 3 transcription support
Adds a VertexAIAudioTranscriptionConfig wired through ProviderConfigManager so
vertex_ai/chirp_3 works on /v1/audio/transcriptions (sync and async) via the
Speech-to-Text v2 recognize API. Auth reuses the standard Vertex credential
resolution (vertex_project/vertex_location/vertex_credentials or ADC); the
location defaults to the us multi-region since chirp_3 is only served from the
us and eu multi-regions, and non-global locations use the regional
<location>-speech.googleapis.com host. Maps language to languageCodes (auto
language detection by default), joins all result alternatives into the
transcript, and tracks cost from totalBilledDuration with a
vertex_ai/chirp_3 price entry at Google's published $0.016/min.
* fix(vertex_ai): map bare ISO-639-1 language codes to BCP-47 for Speech-to-Text
OpenAI clients send language codes like "en", which Google rejects with 400
("not supported by the model chirp_3 in the location us"); Speech-to-Text
wants region-qualified BCP-47 like "en-US". Adds a shared
normalize_transcription_language_to_bcp47 helper in audio_utils (NVIDIA Riva's
transcription config already hand-rolled the same table privately) that maps
common bare codes and passes region-qualified ones through, and applies it in
the Vertex transcription request. Also narrows the response JSON parse guard
to ValueError.
* fix(vertex_ai): drop zero output_cost_per_second so chirp_3 cost tracking works
cost_per_second prefers output_cost_per_second whenever it is not None, so the
0.0 in the chirp_3 entry priced every transcription at $0.00 instead of using
input_cost_per_second. Remove it from both cost maps and pin the behavior with
a regression test computing 18s of chirp_3 audio to ~$0.0048.
* fix(vertex_ai): validate client-controllable location to prevent SSRF in Speech-to-Text
get_complete_url interpolated vertex_location straight into the request host,
and vertex_location is client-controllable on the proxy (it flows from the
request body and is not on the request-body blocklist). An authenticated caller
could send vertex_location="attacker.example/" to point the host at their own
server, so the proxy would POST the audio plus its admin-minted Google bearer
token and x-goog-user-project header to the attacker, exfiltrating a
cloud-platform-scoped OAuth token minted from the admin's credentials.
Factor the location validation the rest of vertex_ai already applied in
get_vertex_base_url (^[a-z][a-z0-9-]*$ plus the global allowance) into a shared
validate_vertex_location helper in common_utils and call it from both the chat
host builder and the new speech host builder. Invalid locations now raise a 400
VertexAIError instead of building a host. Also reject vertex_project values that
carry URL-structural characters, since it lands in the URL path.
Regression tests assert on the parsed netloc so the security property is pinned:
valid locations always resolve to a *speech.googleapis.com host and injection
inputs are rejected.
* fix(vertex_ai): reject unsupported transcription response_format values instead of silently ignoring
|
||
|
Mateo Wang
|
5f864c83ce
|
chore(lint): zero out crash-class pyright rules and ban new type: ignore comments (#32152)
* fix: zero out crash-class basedpyright rules across litellm/ * feat(lint): add LIT009 banning inert type: ignore comments * docs: require bracketed rule and reason on every suppression * chore(lint): ratchet budgets down and zero crash-class pyright limits * fix: narrow auto router routelayer through a local before calling * test: add regression tests for crash-class fixes * fix: drop dead AZURE_AD_TOKEN lookups and word-bound the type-ignore regex |
||
|
Sameer Kankute
|
a16d9c6f9e
|
test(e2e): add live batches suite across providers and routing scenarios (#30958)
* tests: add e2e tests for spend, budgets and llms * style: make chained comparison of status_code clearer Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * remove e2e_tests folder * test: add spend tracking tests * fix: p0 issues, added types and shared functions for each test suite * style: carry clearer status_code comparison into renamed e2e dir * refactor: migrate to gateway client * fix: add new tests, split gateway * test(e2e): add live batches suite across providers and routing scenarios * test(batches): cover real cost tracking on completed batch retrieve * test(e2e): assert managed vs raw file and batch id shapes per routing scenario * test(e2e): assert full response shape of each batches and files endpoint * test(e2e): only accept transitional statuses for a freshly created batch * test(prompt-factory): make test_convert_url deterministic with a data URL picsum.photos is down (HTTP 522), so test_convert_url failed on every run. Swap the live external image for an inline data: URL and assert the round-trip through convert_url_to_base64 genuinely. A data URL is already inline base64 image data, so convert_url_to_base64 now short-circuits it instead of attempting an impossible HTTP fetch; add a regression for that branch in the mapped image_handling test * fix: pass through async image data urls * fix(image-handling): short-circuit data URLs in async path too Bugbot flagged that convert_url_to_base64 returns data: base64 URLs unchanged but async_convert_url_to_base64 still tried to fetch them, so async OCR flows (Bedrock, Azure) would reject inline images the sync path accepts. Add the same guard to the async function and a regression test that asserts the async path returns the data URL without touching the HTTP client * Fix: openai batches lifecycle * Fix: add e2e azure openai tests * Fix e2e for vertex ai * Add all models for testing * test(managed-files): assert idempotent upsert in store_unified_file_id store_unified_file_id switched from create to upsert to avoid UniqueViolationError when re-storing the same unified_file_id (e.g. batch output files stored before metadata is available). Update the unit test to assert the upsert call and its create payload instead of the removed create call. * test(batches): reconcile vertex_ai native batch-id comment with fallback guard * fix(test-config): keep rust-ocr models in model_list by moving files_settings after it * fix(test-config): move batch models after OCR block to keep merge with internal_staging clean * fix(batches): use '24hrs' completion window and allow managed-files listing with provider filter Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * style: ruff format transformation.py and endpoints.py Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(e2e/batches): set Azure raw_model to gpt-4.1-mini-batch to match deployed model Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(vertex-ai/batches): correct completion_window to 24h per Literal type definition * test(vertex-ai/batches): align completion_window assertion to 24h * fix: update managed file metadata on upsert --------- Co-authored-by: mubashir1osmani <mubashir.osmani777@gmail.com> Co-authored-by: Mateo Wang <277851410+mateo-berri@users.noreply.github.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Cursor Agent <cursoragent@cursor.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> |
||
|
Sameer Kankute
|
a2a951a1e9
|
feat(vertex_ai): pass full imageConfig dict for Gemini image generation (#31811)
* feat(vertex_ai): pass full imageConfig dict for Gemini image generation Support all ImageConfig fields (aspectRatio, imageSize, personGeneration, imageOutputOptions) when calling Vertex AI Gemini image generation endpoints. Previously only aspectRatio and imageSize were extracted; other fields were silently dropped. Co-authored-by: Cursor <cursoragent@cursor.com> * style: ruff format vertex_gemini_transformation Co-authored-by: Cursor <cursoragent@cursor.com> * fix(vertex_ai): warn on non-dict imageConfig instead of silently dropping Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> |
||
|
mubashir1osmani
|
85840aef51
|
fix(vertex_ai/files): single media upload for batch files to fix 499s on large uploads (#31653)
* fix(vertex_ai/files): upload batch files in a single media request to fix 499s on large uploads PR #31036 switched the vertex batch file upload from a single GCS media upload to a chunked resumable session. The resumable path sends the body as many sequential PUTs, each waiting a full round-trip to GCS before the next, so a multi-GB upload accumulates hundreds of round-trips and overruns the client/load-balancer request timeout, surfacing as 499s (client closed connection) on files as small as 500MB. This was a regression from the last-known-good commit, where the upload completed as one continuous request. Revert the batch upload to a single uploadType=media request, but stage the transformed payload to a temp file first so peak memory stays bounded (the goal of the resumable rewrite) without the per-chunk round-trips. The temp file is closed deterministically (TemporaryFile unlinks on close), not left to the GC. The now-unused resumable chunked-upload plumbing is removed. Also swap the per-row transform's stdlib json for orjson (parse + serialize), which is ~4x faster on this hot path; the streaming body now emits compact orjson bytes. The request stays synchronous, so the returned file object is real and POST /v1/batches keeps working immediately against the uploaded object. Tests: single media request carries the whole payload with a real Content-Length (no chunked transfer-encoding); failed upload raises; the staged temp file is closed deterministically; byte-for-byte transform parity. * test(vertex_ai/files): mock single media upload POST instead of removed resumable method test_avertex_batch_prediction patched BaseLLMHTTPHandler._aresumable_chunked_upload, which was removed when the batch jsonl upload moved from a chunked resumable GCS session to a single uploadType=media request. Patch the raw httpx.AsyncClient.post that _astage_and_upload_media issues so the real staging, upload and response transform run while the GCS object response is mocked, and assert the media URL and Content-Type. * fix(vertex_ai/files): forward request timeout to media upload, drop orjson, sort imports Forward the per-request timeout through _stage_and_upload_media / _astage_and_upload_media to the GCS POST. Every other upload branch forwards it; the new media path was dropping it, so a caller-provided timeout was silently ignored (the files path passes 600s by default, but a custom request_timeout would not have reached this upload). Regression test asserts the resolved timeout reaches the request (mutation-verified). Revert the orjson swap in the batch transform: importing orjson at module load in this core-path file broke `import litellm` on environments without orjson (the Windows import test). Back to stdlib json; the upload leg dominates large uploads anyway, so the transform-side win was marginal. Fix import ordering in llm_http_handler.py (I001) introduced by the new imports. * fix(vertex_ai/files): stream batch upload to GCS instead of staging to a temp file Addresses a disk-exhaustion concern: staging the full transformed batch body to a local temp file before the GCS request meant an authenticated user could fill the proxy's temp volume with large concurrent uploads (on top of Starlette's input spool). GCS's simple/media upload accepts chunked transfer-encoding, so stream the transform straight to the single media request instead. Each block is produced on a worker thread (the transform never runs on the event loop) and sent chunked, so the body is neither buffered in memory nor written to disk, and the upload is still one continuous request (no per-chunk round-trips, no 499). Drops the temp-file staging, the tempfile/IO imports, and Content-Length computation. Regression test asserts the upload streams (chunked transfer-encoding, no Content-Length) and creates no temp file; mutation-verified that reintroducing staging fails it. |
||
|
Sameer Kankute
|
2cf565ae28
|
test(batches): add 1:1 test file scaffold for batches component paths (#30529)
* test(batches): add 1:1 test file scaffold for batches component paths Co-authored-by: Cursor <cursoragent@cursor.com> * Add harness test for create batch endpoint * Add retrieve endpoint harness tests * Add list endpoint harness tests * Add cancel endpoint harness tests * Add cancel endpoint harness tests * Add test for litellm/batches/main.py * Add test for litellm/tests/test_litellm/batches/test_batch_utils.py * Add handler and transformation tests for all providers * Fix: run batches tests in cicd * fix(tests): remove azure/__init__.py that shadowed azure namespace package Adding __init__.py to tests/test_litellm/llms/azure/ caused pytest to insert tests/test_litellm/llms/ into sys.path[0], making our empty azure/ dir shadow the real azure-identity namespace package. Any test that patched azure.identity.* would then fail with AttributeError. * style(tests): apply ruff format to test_batch_utils.py Base migrated the formatter from black to ruff format (#31317); reformat the batches scaffold test file to match. --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Shivam Rawat
|
d515e5bf05
|
fix(vertex_ai): append rawPredict suffix for custom api_base on /v1/messages (#31529) | ||
|
Sameer Kankute
|
4476923ac4
|
test: add realtime proxy e2e suite across providers (#30960)
* tests: add e2e tests for spend, budgets and llms * style: make chained comparison of status_code clearer Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * remove e2e_tests folder * test: add spend tracking tests * test: multi-window budgets coverage * fix: p0 issues, added types and shared functions for each test suite * chore: add config.yml * test: passthrough endpoints stream/non-stream e2e * style: carry clearer status_code comparison into renamed e2e dir * fix: rename cost breakdown function * fix: pydantic validation for budget info, dont allow explicit type cast * refactor: migrate to gateway client * test: add custom pricing tests * chore: change master key * test(e2e): address greptile review feedback Remove the duplicate cache/cache_params block in the gateway config so the two can't silently diverge under future edits. Reorder the soft-budget test to assert the call isn't a budget block before require_successful_call, since that helper hard-fails any non-2xx and left the budget-block check unreachable; the misleading "skip" comment is corrected. Add a deferred delete in test_budget_delete_removes_it so a failed delete doesn't leak a budget on the shared proxy. Scope the spend_tracking sys.path insertion in pytest_sessionfinish to just the cleanup import so a broader "pytest tests/" run isn't left with a mutated path. * test(e2e): drop misleading skip comment on require_successful_call require_successful_call fails hard, it does not skip; the trailing comment was factually wrong. The function name already states intent, so the comment is removed in both per-model and tag budget helpers. * test(e2e): assert budget-isolation invariant before success check On the should-still-succeed path of the per-model and tag isolation tests, check is_budget_block before require_successful_call. If the isolation bug fires the unaffected model/tag is blocked, so asserting the specific 'blocked by X' invariant first yields the diagnostic message instead of a generic upstream-failure. Matches the ordering in test_soft_budget_e2e.py. * fix(e2e): guard spend-log truncate on skip and stop returning unrelated priced rows * fix(e2e): run case init() inside try so partial-init failures tear down run_case called case.init() outside the try/finally that runs teardown(), so a case that registers cleanups progressively (create team, then user, then key) and then fails partway through init() would leak the already-created entities on the long-lived shared proxy. Move init() inside the try so teardown always runs. Add a regression test that registers a cleanup then raises mid-init and asserts the resource is still released. * test(e2e): mark known pricing-leak isolation test xfail(strict) test_custom_pricing_is_isolated_from_sibling_deployment documents a real proxy gap (a deployment's custom per-token pricing leaks into the shared cost map for sibling deployments of the same underlying model) and was left unconditionally failing, which pollutes the suite's pass/fail signal. Mark it xfail(strict=True) so the suite stays green while the leak persists and turns into a failure the moment isolation is fixed, prompting the marker's removal. * refactor(e2e): make suite pass its shipped strict basedpyright config The suite ships tests/pyrightconfig.json (strict, no Any), but basedpyright --project tests reported four errors in it: three reportAny on the parametrize ids=lambda c: c.__name__, and one reportUnusedFunction on the underscore-prefixed autouse fixture _require_live_proxy. Replace the untyped lambda with a typed _case_id(case_cls: Type[_BudgetCase]) -> str so the ids are no longer Any, and rename the fixture to require_live_proxy so basedpyright no longer treats it as an unused private function (it is referenced only by pytest's autouse machinery). basedpyright --project tests now reports zero errors. * fix(tests/e2e): gate spend-log truncate on e2e marker, not test directory * test(e2e): run harness unit tests without a live proxy The autouse session fixture skipped the whole tests/e2e session when no proxy answered, which also skipped test_lifecycle.py, a pure unit test of run_case that never touches the proxy. A regression test that silently skips gives no signal, so the skip now lives in pytest_runtest_setup gated on the same e2e marker the spend-log truncate guard already uses: live tests skip when no proxy is up while harness unit coverage always runs. The liveness probe is cached with lru_cache so it still runs once per session * test(e2e): clean up gateway config comment debris Fix the typo on the header comment and drop the orphaned namespace/ttl comment remnants left indented under cache_params; the active values are already set above. Flagged by greptile review. * fix: add new tests, split gateway * test(e2e): type the redis spend-counter probe for strict basedpyright The new cold-counter reseed test drove its redis client untyped, so the strict tests/pyrightconfig.json (reportUnknown*, reportAny) flagged ten errors once the file landed: scan_iter/get came back unknown and the pool.map lambda had an untyped parameter. Annotate the client as redis.Redis[str] via a TYPE_CHECKING import (the runtime import stays lazy so the suite still skips, not errors, when redis is absent), which resolves scan_iter to Iterator[str] and get to str | None, and replace the lambda with a typed inner function mirroring _burst. basedpyright --project tests is back to zero errors. * test(e2e): xfail the known team multi-window failure and isolate member teardown Greptile flagged two issues in the mirrored split-gateway commit. The team multi-window budget test documents a real /team/new write bug (budget_limits go straight to the Json? column and Prisma 500s, unlike the json.dumps'd key and /team/update paths) and was left as an unconditional hard failure, which would turn any live-proxy CI run red; mark it xfail(strict=True) like the custom-pricing isolation test so the suite stays green while the bug persists and flips to a failure the moment the write is fixed and the marker should go. The class-scoped member fixture in test_team_member_budget_e2e.py tore down its key, user, and team sequentially with no exception isolation, so a failed delete_key would strand the user and team on the long-lived shared proxy. Route cleanup through a ResourceManager: register each delete progressively and run them LIFO best-effort in a finally, so a partial-setup failure still releases what came before and one failed delete never blocks the rest. * test: add realtime proxy e2e suite across providers Add tests/realtime_e2e covering the proxy realtime websocket endpoint end to end against live providers (openai, azure, gemini, vertex_ai, bedrock, xai). Two layers: a raw-websocket suite asserting the normalized OpenAI GA event sequence, delta/transcript consistency, usage, and a full tool-call round-trip; and a pipecat smoke driving the proxy through the GA OpenAIRealtimeLLMService. Tests carry a new realtime_e2e marker and skip cleanly when the proxy or provider creds are absent, so they stay out of the default unit run. * test: move realtime e2e suite into tests/e2e harness Replace the standalone tests/realtime_e2e with a tests/e2e/realtime suite that follows the existing e2e conventions: a session-scoped client fixture, a frozen-dataclass RealtimeClient wrapping the shared Gateway, pydantic models for every sent and received event, and the e2e marker with the parent harness's liveness skip. The suite opens the proxy realtime websocket (websockets.sync to stay synchronous like the rest of the harness) and asserts the normalized OpenAI GA event sequence for a text conversation plus a full tool-call round-trip, parametrized across providers. A provider whose realtime alias is not configured on the proxy skips via /model/info. Adds a gemini realtime model to the gateway config and fixes the openai realtime model id. * test: add pipecat realism layer to realtime e2e suite Add test_realtime_pipecat_e2e driving the same providers through pipecat's GA OpenAIRealtimeLLMService with base_url pointed at the proxy, as a coarse realism check on top of the raw-websocket suite. Each test stays synchronous and runs the async pipecat pipeline via asyncio.run, and the module skips unless pipecat-ai is installed. Lift the shared provider matrix, ws-url helper, and skip helper into realtime_client so both suites use them. * fix(e2e): parse GA realtime transcript events in e2e client The realtime e2e client speaks the GA protocol, but transcript() only aggregated beta delta event names. Handle GA deltas, fall back to response.done output, and accept nested usage details on response.done. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(e2e): address realtime code-review findings - Use the real openai/gpt-4o-realtime-preview model ID in the gateway config (gpt-realtime-2 does not exist and would fail every live test) - Pass a bare base_url to pipecat's OpenAIRealtimeLLMService so pipecat can append ?model= itself; the previous realtime_ws_url already contained ?model= causing a malformed duplicated query parameter - Wrap connection.recv() in a try/except TimeoutError in collect_until so a deadline expiry inside recv preserves the collected-events diagnostic instead of raising a bare, message-free exception Co-authored-by: Cursor <cursoragent@cursor.com> * fix(e2e): filter configured_models to mode:realtime entries only ModelInfoEntry.model_info used CustomPricing (extra="ignore") so the mode field from /model/info was silently dropped, making it impossible to distinguish realtime from non-realtime deployments. Add an optional mode field to CustomPricing and filter configured_models() to entries whose model_info.mode == "realtime" so skip_if_unconfigured never accidentally skips a realtime test due to a naming-pattern collision with a non-realtime deployment. Co-authored-by: Cursor <cursoragent@cursor.com> * Update litellm-config.yml * fix(e2e): use TypeVar instead of PEP 695 generic in realtime parse_last PEP 695 type-parameter syntax (def f[T: Bound](...)) is only parseable on Python 3.12+, but the project declares requires-python >=3.10. Importing the realtime e2e client on 3.10/3.11 raised a SyntaxError before any test could run. Switch parse_last to the backport-safe TypeVar idiom so the suite imports across the full supported range. * fix(e2e/realtime): use GA openai/gpt-realtime model id The realtime gateway config used openai/gpt-realtime-2, which is not a real OpenAI model id and would 404 once live OpenAI realtime credentials are wired in. The GA speech-to-speech model is openai/gpt-realtime (snapshot gpt-realtime-2025-08-28); switch the openai-realtime alias to it. * fix(realtime): harden Gemini/Vertex Live for audio-native e2e Coerce TEXT responseModalities to AUDIO on native-audio and flash-live models, suppress the orphan turnComplete response.done that arrives immediately after tool results, omit function_response.id on Vertex, stop appending client query params to Gemini/Vertex WSS URLs, and add regression tests for these paths. Co-authored-by: Cursor <cursoragent@cursor.com> * Add xai full compatibility * Add working vertex ai realtime tests * Add audio + server vad e2e tests * Add config for e2e testing models * Add fix xai server vad * fix: use correct OpenAI realtime model ID in e2e gateway config openai/gpt-realtime is not a valid model; replace with the correct openai/gpt-4o-realtime-preview model ID to prevent model-not-found errors when running the openai-realtime e2e tests. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * revert: restore openai/gpt-realtime model ID gpt-realtime is a valid model; reverting the unnecessary change. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: resolve UP006 violations, mock test failures, and stale spec field - Guard gemini setup-without-tools deferral with litellm.gemini_live_defer_setup flag so the default (False) path sends setup immediately, fixing two failing mock tests: test_client_ack_caches_setup_to_prevent_duplicate_session_update_setup and test_deferred_setup_sends_session_update_before_buffered_audio - Replace deprecated typing generics (Dict, List, Tuple, Optional) with builtin equivalents in xai/realtime/transformation.py, gemini/realtime/transformation.py, and realtime_streaming.py to satisfy the UP006 ruff-strict ceiling - Remove 'role' from OpenAPI compliance test expected fields; Google removed it from the Interaction schema in their live spec Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: use Optional[dict] in xai normalizer to preserve Black line-split dict[str, Any] | None is shorter than Optional[Dict[str, Any]] by enough that Black collapses the _normalize_usage signature to a single line (86 chars), conflicting with the existing multiline format. Using Optional[dict[str, Any]] keeps the line at 90 chars (> 88 limit) so Black preserves the multiline shape, while still satisfying UP006 by replacing Dict with dict. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: remove proxy-level setup-tools deferral, delegate to transformer The _gemini_setup_deferred / _gemini_pre_setup_buffer block in _send_to_backend was double-deferring: GeminiRealtimeConfig already handles the session.update-to-setup mapping internally and always returns a ready-to-send setup on the first session.update call (session_configuration_request=None). The proxy layer was incorrectly holding back that setup waiting for tools that the transformer had already incorporated. Removing the block fixes two failing tests: test_client_ack_caches_setup_to_prevent_duplicate_session_update_setup test_deferred_setup_sends_session_update_before_buffered_audio Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * refactor: abstract Gemini protocol keys out of core and use cost map for live model detection Move Gemini-specific message key knowledge (setup, realtimeInput, clientContent, toolResponse) out of the core RealTimeStreaming module into provider-level methods. BaseRealtimeConfig gains is_setup_message and is_content_message (both default False); GeminiRealtimeConfig overrides them with the actual Gemini key checks. Add gemini_native_audio and gemini_audio_only_live capability flags to the 10 affected model entries in the cost map. _is_audio_only_live_model and _is_native_audio_model now read from the cost map first and fall back to the existing string markers for models not in the map. * fix: apply black formatting and register gemini capability fields in schema * refactor: drop string-marker fallback; resolve audio-only live models via cost map only * fix: use registered cost-map model name in vertex realtime tests * fix: patch cost map in tests so they don't depend on remote main branch state * fix: align gateway config vertex-realtime model ID with cost-map registered name * fix: patch gemini-2.5-flash-native-audio in cost map fixture for CI * fix(e2e): use correct OpenAI realtime model id in gateway config Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(e2e): add budget rescheduler short intervals to gateway config Without proxy_budget_rescheduler_min/max_time set, the rescheduler defaults to ~600s, causing all budget-reset e2e tests to timeout before the reset fires. Set to 5–10s so tests complete within 90s. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore(e2e): strip non-realtime files from PR scope Restore budget, spend-tracking, and custom-pricing test files to their litellm_internal_staging state. Keep the mode field addition to CustomPricing in models.py (needed by realtime configured_models filter). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(tests): restore async_realtime regression test and add missing fixture - Restore the end-to-end async_realtime regression test for Vertex query-param forwarding; the previous unit-only version did not exercise the code path where the original bug lived - Add patch_gemini_audio_cost_map_entries fixture to test_gemini_audio_only_live_models_drop_text_from_text_audio_combo so it does not depend on the cost map having gemini_audio_only_live set in CI Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(lint): resolve ANN401 violations in realtime streaming code Define RealtimeEventNormalizer Protocol and replace bare Any annotations with typed alternatives (object for event/value params, the Protocol for the normalizer) to stay within the strict-rule budget. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * style: black format realtime_streaming.py * fix(tests): add gemini_native_audio and gemini_audio_only_live to model prices schema * fix(lint): fix I001 import sort order in realtime_streaming.py * fix(lint): restore import litellm to correct position before from-litellm imports * undo budget removal * test(e2e): pin explicit credentials for gemini and vertex realtime models * test(e2e): share keepalive-safe LiteLLMRealtimeLLMService across pipecat suites The pipecat smoke test drove the proxy through the stock OpenAIRealtimeLLMService, which sends websocket keepalive pings at its default interval. The proxy does not answer them, so the connection is closed with a 1011 before the run completes. Move the proxy-aware LiteLLMRealtimeLLMService (keepalive disabled) into a shared pipecat_service module and use it from both the smoke and audio suites. * test(e2e): document that LiteLLMRealtimeLLMService._connect keeps the ?model= param The proxy routes realtime websockets on the ?model= query param, and pipecat's OpenAIRealtimeLLMService.__init__ bakes it into self.base_url before _connect runs. Passing self.base_url through preserves it; spell that out so the override is not misread as dropping the param. * fix(realtime): set _content_sent_after_setup only after the backend send succeeds A failed content send used to flip _content_sent_after_setup to True before the send was confirmed, mirroring the correct-on-failure ordering the adjacent session-config cache already follows. If the send raised, the flag stayed True and a later session.update that produced a setup frame was silently dropped even though the backend never received any content. Set the flag after the send succeeds and add a regression test that fails if the ordering is reverted. * fix: normalize realtime passthrough events * refactor(realtime): declare patch_outgoing_session on normalizer Protocol; fix wav chunk return type The RealtimeEventNormalizer Protocol only declared should_drop and normalize, so the outgoing session.update patch went through a getattr(..., None) lookup even though should_drop/normalize are called directly. The sole implementer (XAIRealtimeNormalizer) already provides patch_outgoing_session, so declare it on the Protocol and call it directly for consistent, fully-typed dispatch. Also correct _load_wav_chunks' return annotation from list[bytes] to tuple[list[bytes], int]; it returns (chunks, sample_rate) and the caller unpacks both. --------- Co-authored-by: mubashir1osmani <mubashir.osmani777@gmail.com> Co-authored-by: Mateo Wang <277851410+mateo-berri@users.noreply.github.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Cursor Agent <cursoragent@cursor.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> |
||
|
Yassin Kortam
|
29c254d3d3
|
fix(vertex): stop O(n^2) re-parse of accumulated Gemini stream JSON (#31297)
handle_accumulated_json_chunk re-ran json.loads on the entire accumulated buffer after every fragment. For a streaming response fragmented across many chunks that is O(n^2) total work in a single GIL-holding C call, so a large enough Gemini response freezes the asyncio event loop for seconds, liveness probes time out, and the proxy pod gets killed and restarted. A complete Gemini stream value is a JSON object or array, so the buffer can only become parseable once its last non-whitespace byte can close one. Gate the json.loads attempt on that, which makes the common fragmented-response case parse roughly once instead of once per fragment. An 8MB payload drops from a 6.9s event-loop freeze to ~0.3s with identical parsed output. Resolves LIT-3503 Fixes #26181 |
||
|
Sameer Kankute
|
062d8ceeed
|
fix(vertex_ai): prevent stale Vertex bearer token causing /v1/messages 401 after token expiry (#31276)
* fix(vertex_ai): prevent stale Vertex bearer token causing /v1/messages 401 after token expiry Router shallow-copies litellm_params so extra_headers is a shared reference. The chat/completions path was calling headers.update() on that shared dict, persisting the Vertex OAuth bearer. After ~1 h the token expired and /v1/messages kept reusing it (skipping refresh due to Authorization-already-present guard). - Build a new headers dict in the Claude partner-models completion path instead of mutating the shared extra_headers object. - Always call _ensure_access_token() in validate_anthropic_messages_environment regardless of an existing Authorization header; the token cache makes this cheap. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(vertex_ai): copy headers in validate_anthropic_messages_environment to prevent shared-dict mutation Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> |
||
|
Sameer Kankute
|
7eacdd5258
|
chore: litellm oss staging 250626 (#31305)
* fix(anthropic): support Bearer auth for custom api_base endpoints (Fixes #30926) * style: format common_utils.py with black * fix(anthropic): extract api_base from litellm_params in batches/files validate_environment * fix(anthropic): scope Bearer key check to custom api_base endpoints * fix(streaming): reset Anthropic message_start cursor (output_tokens=1) when no message_delta arrives The Anthropic streaming protocol emits `message_start.usage.output_tokens=1` as a placeholder cursor; the real cumulative output count only arrives in the final `message_delta` event. When a stream is cancelled before `message_delta` lands (common for thinking models on long-tail prompts), ChunkProcessor._calculate_usage_per_chunk's last-wins accumulator left completion_tokens stuck at 1. Because 1 is truthy, the `completion_tokens or token_counter(text=...)` fallback in calculate_usage() never fired, and requests were billed for 1 output token even when several thousand tokens of text had actually streamed. Fix: track whether any chunk's completion_tokens exceeded 1 (saw_non_cursor_completion). If the only update we saw was the cursor, reset completion_tokens to 0 so the text-based fallback estimates from the real completion content. Legitimate 1-token completions (model returns "Yes." etc.) are unaffected in practice — token_counter on a 1-token completion_output also yields ~1, so billing stays approximately correct. Tests: - TestAnthropicCursorBug (6 cases) — pins the post-fix behavior - TestNonAnthropicStreamingIntact (2 cases) — guards against regression on providers without the cursor pattern All 8 new tests pass; 9 existing streaming_chunk_builder_utils tests still pass. * fix(streaming): scope cursor reset to anthropic provider + recognize message_delta arrival Addresses both Greptile P2 threads on PR #30420: CLASS A — Anthropic-specific heuristic was applied globally ============================================================ The `completion_tokens == 1 and not saw_non_cursor_completion` reset lived in provider-neutral `streaming_chunk_builder_utils.py`. Any non-Anthropic provider that legitimately reports completion_tokens=1 in a single usage chunk (perfectly normal for short OpenAI / Bedrock / Vertex single-token replies with stream_options.include_usage=true) would have its value silently rewritten to 0 and re-billed via token_counter — producing a different number than what the provider actually charged. Fix: gate the reset on `custom_llm_provider == "anthropic"`, resolved from the first chunk's `_hidden_params` (the same field set by streaming_handler.py:722 on the live path). Unknown / missing provider is treated as non-Anthropic and skips the reset, so newer providers and custom plugins are also safe by default. CLASS B — `saw_non_cursor_completion` missed legitimate single-token replies ============================================================ Previous condition was `usage_chunk_dict["completion_tokens"] > 1`, which never fires for an Anthropic stream where the model legitimately emits exactly one output token (e.g., "Yes."). Anthropic still sends message_start (output_tokens=1, the cursor) AND message_delta (output_tokens=1, the real value) — same value, but two distinct usage events. The old check couldn't tell that apart from a cancelled stream where only message_start landed. Fix: track `completion_usage_updates` and flip `saw_non_cursor_completion` when EITHER (1) the value exceeds 1 (definitely not a placeholder), OR (2) we've seen >=2 completion-bearing usage events (positive evidence that message_delta arrived). Cancelled cursor-only streams still have exactly one event and still hit the reset; cache chunks with completion_tokens=0 don't count toward the threshold. Tests ============================================================ - _make_chunk now sets `_hidden_params["custom_llm_provider"]` (default "anthropic") so the gate is exercised by every existing test — none of them needed assertion changes besides the legitimate-single- token case, which now expects exactly 1 (was a fuzzy 0..3 range). - New: test_anthropic_cache_only_chunks_after_message_start_still_resets - New: test_non_anthropic_provider_completion_tokens_one_not_reset - New: test_unknown_provider_completion_tokens_one_not_reset 11/11 tests pass. * chore: add Co-authored-by trailer for attribution Co-authored-by: songkuan-zheng <songkuan-zheng@users.noreply.github.com> * fix(anthropic): preserve messages cache usage * style(anthropic): format messages cache usage helper * fix(anthropic): accept integral float cache token counts Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(anthropic): accept integral float cache token counts * test(anthropic): cover cache usage edge cases * fix(gemini): preserve thoughtSignature for server-side tool responses When Gemini API returns toolCall and toolResponse parts, they might have different thoughtSignatures. Previously, LiteLLM merged them into a single dict, overwriting the response's thoughtSignature with the call's. This fix extracts them separately and re-injects them correctly. TAG=agy CONV=755b21d0-3200-40bc-bd1a-bb58a378a9a6 * fix(gemini): address PR comments on thoughtSignature handling - Fix orphan-response thoughtSignature regression by copying thought_signature to response_thought_signature - Add missing assertions in existing tests - Add new unit tests for orphan-response signature handling TAG=agy CONV=755b21d0-3200-40bc-bd1a-bb58a378a9a6 * feat(mcp): include server alias and server_id in mcp_info response - Add alias and server_id fields to mcp_info object in /mcp-rest/tools/list endpoint - Update rest_endpoints.py to surface alias from server config - Add test coverage in test_mcp_server.py and test_rest_endpoints.py Fixes #31015 * fix(proxy): reject non-finite spend via validate_finite_spend A NaN/-inf spend would bypass spend >= max_budget enforcement. Add a shared finite-value guard, defined above the litellm.proxy.* imports to avoid the module-level cyclic-import warning. * fix(proxy): require admin for any /key/update spend, reject non-finite Gate the admin check on the presence of `spend` (not a value diff): the DB spend lags the live cross-pod counter, so an "unchanged" spend on the non-admin path let a key owner / team member overwrite the live counter below real usage. Also reject NaN/+-inf spend before the DB write. * fix(proxy): invalidate spend counter on /user/update spend change A direct spend change on /user/update wrote the DB row but left the warm cross-pod counter at the stale value, so enforcement kept reading the old spend. Invalidate spend:user:{user_id} after the write (reseed-from-DB), and reject non-finite spend before the write. * fix(cache): route Bedrock semantic-cache sync embedding through the Router (#28244) The semantic cache's embedding model is a proxy Router alias whose AWS credentials (aws_role_name, aws_session_name) live only in the Router deployment's litellm_params. The sync embedding paths called litellm.embedding() directly, bypassing the Router, so they could neither resolve the alias nor assume the configured role; cross-account Bedrock semantic caching failed with "bedrock:InvokeModel is not authorized". On Redis this surfaced at proxy startup because redisvl's CustomTextVectorizer eagerly fires a dimension-probe embedding during cache construction, while llm_router is still None. Fix A: make the sync paths mirror the already-correct async paths. A shared, dependency-injected helper (litellm/caching/_embedding_router.py) decides whether to route through llm_router.embedding(...) when the model is a Router deployment, else fall back to direct litellm.embedding(...). Redis and qdrant sync set_cache/get_cache now precompute the embedding and pass vector= to the backend, exactly as the async astore/acheck already do. Both async _get_async_embedding methods are unified onto the same helper and now forward the caller's full metadata instead of a hand-picked subset. Fix B (Redis only): defer redisvl index construction from __init__ into a lazy, memoized llmcache property, so the dimension-probe embedding fires on first cache use, after llm_router is wired. A failed build is not memoized, so a transient outage recovers on the next request. Known limitation: resolve_embedding_router gates on an exact model-name match (same as the shipped async path); wildcard/alias/team-public routes still fall back to direct embedding. Tracked as a follow-up. * fix(cache): harden embedding-router and shrink Any surface (review) Address review feedback on the semantic-cache aws-role fix (#28244): - resolve_embedding_router now skips deployment entries missing model_name instead of raising KeyError on a malformed model_list (Greptile P2); add a regression test that fails on the old direct-key access. - Replace the `**kwargs: Any` passthrough on the four cache _get_embedding / _get_async_embedding helpers with an explicit, typed `metadata: Optional[Dict[str, Any]] = None` parameter. The helpers only ever consumed kwargs["metadata"], so this is behavior-preserving, makes the forwarded field obvious at the call site, and removes three bare-Any annotations (keeps the strict-rule ANN401 budget within ceiling). - Note in _build_llmcache that redisvl's dimension-probe embedding adds one extra billable embedding on the first cache request (Greptile P2). * fix(bedrock_mantle): correct responses routing for openai.gpt-5.x models Dashboard Test Connection for bedrock_mantle/openai.gpt-5.4 and openai.gpt-5.5 was failing with maximum recursion depth errors and "model does not exist" Route detection in the bedrock provider matched route tokens by plain substring, so the bedrock_mantle/ prefix was mistaken for the mantle/ invoke route and the body model was rewritten to bedrock_openai.gpt-5.5; route tokens now only match at a path-segment boundary so the bare model name is preserved A responses-mode model whose provider has no responses config bounced forever between the responses API and chat completions; the responses to completion fallback now tags its call so completion() does not bridge back, breaking the loop The Test Connection endpoint hardcoded the test mode to chat, which disabled mode auto-detection for responses-only models; the default is now None so the mode is detected from model capabilities acompletion() now drops a duplicate acompletion kwarg before building the partial and treats model_info=None as an empty dict to avoid a NoneType crash * test(bedrock_mantle): cover route guard and bridge flag; fix reportArgumentType regression Adds the regression coverage codecov flagged on the two responses to completion bridge guard lines and the bedrock route-prefix helper. The handler tests drive both the sync and async fallback paths with litellm.completion and litellm.acompletion mocked, and assert the forwarded kwargs carry _skip_responses_api_bridge=True, so dropping either flag line fails the suite. The common_utils tests assert that bedrock_mantle/openai.gpt-5.x no longer resolves to the mantle route while the genuine mantle/ and bedrock/mantle/ ids still do, exercising both branches of _model_has_route_prefix. Also aligns update_messages_with_model_file_ids model_id to Optional[str], matching its Responses API sibling, so the defensive model_info fallback no longer introduces a new reportArgumentType in completion(); the file-id lookup narrows model_id before the dict get * chore(ui): sync generated OpenAPI types for optional test_connection mode The test_model_connection mode body param default changed from chat to None so the mode is auto-detected from model capabilities, which makes the field optional in the proxy OpenAPI spec. Regenerate the committed schema so the dashboard types match: mode becomes optional and the description and default JSDoc follow the spec, keeping the Check UI API Types Sync gate green * refactor(bedrock): match all explicit route prefixes at path-segment boundary Migrates the remaining substring route checks to the existing _model_has_route_prefix helper so every explicit route token matches only as a leading path segment, consistent with get_bedrock_route and the mantle route. Covers _explicit_converse_route, _explicit_claude_platform_route, _explicit_invoke_route, _explicit_agent_route, _explicit_agentcore_route, _explicit_converse_like_route, _explicit_async_invoke_route and _explicit_openai_route. This also stops invoke/ from substring-matching async_invoke/. Route precedence and order are unchanged, and a note on the segment invariant is added to the helper docstring * test(bedrock): cover explicit route prefix segment matching Exercises all eight migrated _explicit_*_route helpers (converse, converse_like, invoke, async_invoke, agent, agentcore, claude_platform, openai) directly: each matches its token as a leading path segment and rejects the token glued to a preceding segment, so reverting any method to the old substring check fails the suite. Also asserts invoke/ no longer matches async_invoke/ models, the concrete improvement of the segment-boundary migration * test(proxy): assert negative spend is allowed (one-time grant use-case) Negative spend is intentionally permitted so admins can grant extra allowance for the current budget period only, without raising the recurring budget ceiling. Cover it explicitly in validate_finite_spend and via the /user/update invalidation test. * fix(google_genai): forward native generateContent top-level fields Google's native generateContent REST body carries safetySettings, toolConfig, cachedContent and labels at the top level as siblings of generationConfig. The proxy's :generateContent endpoint spread them into agenerate_content as loose kwargs and then dropped them, so callers had to wrap them in extra_body for them to take effect; safetySettings, for instance, was silently ignored The provider config now exposes the native top-level field names and setup_generate_content_call collects whichever are present, merging them into the outgoing request body through the existing extra_body merge so they reach Google verbatim. An explicit extra_body still wins on conflict. The sync generate_content_stream path now also forwards systemInstruction, matching the other three entry points Fixes #12671 Claude-Session: https://claude.ai/code/session_016MFtMXokCjT8u6mvyASudK * fix(proxy): resolve env refs for DB-stored models * fix(proxy): restrict DB env ref resolution * fix(proxy): block team DB env ref resolution * fix(lint): resolve ANN401/UP045/C901 strict-gate violations - Replace Optional[X] with X | None (UP045) in 8 files - Replace Any return/param types with concrete types or object (ANN401) - Extract _make_api_key_auth_header helper to reduce get_anthropic_headers complexity below C901 threshold (17 → 14) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(anthropic): preserve x-api-key for custom endpoints; opt-in Bearer via prefix Users who pass a key already prefixed with "Bearer " get Authorization: Bearer. All other keys continue to use x-api-key, preserving backward compatibility with custom api_base endpoints that expect x-api-key rather than Authorization. Also consolidates get_auth_header to reuse _make_api_key_auth_header helper, eliminating the duplicated custom-endpoint routing logic. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * revert(anthropic): restore Bearer routing for non-sk-ant- keys on custom api_base The backwards-compat change broke existing tests that verify the intentional Bearer-for-custom-base behavior (Fixes #30926). Restore original logic while keeping the _make_api_key_auth_header helper for code deduplication. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(anthropic): gate Bearer-for-custom-base behind use_bearer_for_custom_base flag Previously the auth-header switch from x-api-key to Authorization: Bearer applied unconditionally for non-sk-ant- keys on a custom api_base, silently breaking existing deployments that proxied to gateways expecting x-api-key. Introduce use_bearer_for_custom_base: bool = False on _make_api_key_auth_header, get_anthropic_headers, and get_auth_header. validate_environment reads it from litellm_params so callers can opt in per-model without any API surface change. Tests updated to pass use_bearer_for_custom_base=True where Bearer behavior is asserted. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(redis): apply namespace prefix in delete_cache and async_delete_cache (#29981) DEL was the only Redis cache operation that skipped check_and_fix_namespace, so it targeted the raw SHA256 hash (e.g. 3997c4...) rather than the namespaced key (litellm:3997c4...). This caused two problems: a Redis NOPERM error on deployments with an ACL restricting DEL to the litellm:* pattern, and a silent no-op on all other deployments since the un-prefixed key was never stored. * style(anthropic): reformat common_utils.py with Black (--target-version py312) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: preserve cache metadata and spend counters * style: apply ruff format to streaming_iterator.py * refactor: reduce complexity of usage/spend helpers to satisfy strict ruff gate Extract Anthropic message_start cursor reset into _reset_anthropic_cursor_completion_tokens and the cross-pod spend-counter invalidation into _invalidate_user_spend_counter_if_changed, keeping both _calculate_usage_per_chunk and _update_single_user_helper under the max-complexity ceiling. Use builtin generics in the new signatures so no new UP006 violations are introduced. Behavior unchanged. --------- Co-authored-by: rupak-eng <rupakji99@gmail.com> Co-authored-by: songkuan-zheng <252822057+songkuan-zheng@users.noreply.github.com> Co-authored-by: songkuan-zheng <songkuan-zheng@users.noreply.github.com> Co-authored-by: Kannan Priyadharshan <kpd2204@gmail.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Marco Georgaklis <mgeorgaklis@google.com> Co-authored-by: Anjaiah Methuku <anjaiahspr@gmail.com> Co-authored-by: Andrii Butko <booandrew23@gmail.com> Co-authored-by: Kent <kingdooo@gmail.com> Co-authored-by: kunal2002 <k.nayyar2002@gmail.com> Co-authored-by: Ali Khan <alirazakhan.offi@gmail.com> Co-authored-by: jesco-absolut <team@srswti.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> Co-authored-by: Matt Hill <mhill@dataminr.com> Co-authored-by: Cursor Agent <cursoragent@cursor.com> |
||
|
Mateo Wang
|
6e3540856c
|
fix(vertex): preserve Gemini Embedding 2 usageMetadata for cost tracking (#31354)
* fix(vertex): preserve Gemini Embedding 2 usageMetadata for cost tracking * style(vertex): apply ruff format to batch_embed_content_transformation * fix(vertex): bill files/ image refs in Gemini embedContent at per-image rate Resolved files/... references whose mime type is an image were not detected by _is_image_element, so image_count stayed 0 and generic_cost_per_token fell back to the text token rate instead of input_cost_per_image. Thread the resolved_files mapping into the usage builder so resolved image references are counted and billed per image. Also modernize the _flatten_input return annotation to satisfy the ruff UP006 strict gate. * fix(vertex): bill Gemini embedding audio per-second and stop video+audio double-billing Audio-only embedContent responses set audio_tokens, but generic_cost_per_token only charges audio via input_cost_per_audio_token. gemini-embedding-2 prices audio via input_cost_per_audio_per_second, so spend stayed at $0. Plumb a new audio_length_seconds field through PromptTokensDetailsWrapper, parse it in _parse_prompt_tokens_details, and bill it from _calculate_input_cost. The vertex embedding transformation derives audio_length_seconds from audio_tokens using the documented 32 tokens/sec Gemini rate. The 1-token text floor that protects video billing only fired when no other modality was billable, but audio presence flipped that flag, leaving text_tokens at zero for video+audio responses. generic_cost_per_token then rewrote text_tokens to prompt_tokens minus audio_tokens (the video token count), charging video tokens as text on top of the per-second video cost. The rewrite trigger is text_tokens == 0 and image_count == 0; align the floor with that trigger and ignore audio_tokens. --------- Co-authored-by: Cursor Agent <cursoragent@cursor.com> |
||
|
mubashir1osmani
|
56825926af
|
fix(vertex/files): stream OpenAI->Vertex batch JSONL uploads (#31036)
* fix(vertex/files): stream OpenAI->Vertex batch JSONL uploads to fix OOM on large files
Large (1GB+) batch JSONL uploads to Vertex AI / GCS caused OOM or killed the worker
because the request body was buffered and multiplied 2-3x in size. The create-file
path is now streaming end-to-end: transform_create_file_request returns a
ResumableChunkedUploadConfig carrying a lazy _OpenAIToVertexBatchUploadStream, and the
HTTP handler opens a GCS resumable session and PUTs the body in bounded 8 MiB chunks
(Content-Range, 308 between chunks) so the transformed payload is never held in full.
The proxy /v1/files endpoint streams from Starlette's spooled upload handle instead of
reading the whole body, and batch rate limiting counts tokens and models in a single
streaming pass.
Only gcs_bucket_name is supported for the GCS target; the legacy bucket_name key is
intentionally not read.
Also removes the unreachable VertexAIFilesHandler create path and everything only it
kept alive (VertexAIJsonlFilesTransformation, _stream_openai_jsonl_to_vertex, the legacy
transform helpers), plus the orphaned batch_utils helpers the streaming rewrite replaced.
* fix(batches): return original JSONL on unparseable row to avoid silent batch truncation
The streaming rewrite of replace_model_in_jsonl accumulated physical lines and
skipped a row on JSONDecodeError to support multi-line objects, but a genuinely
malformed or truncated row never completes: it poisons the buffer, swallows every
following row, and the function still returned the partial rewrite (the rows before
the bad one, already model-rewritten) as if the batch were complete. That turned the
pre-rewrite behavior of returning the original file unchanged (so the provider rejects
the bad batch loudly) into a silent partial submission.
Restore the original-content fallback: when an unparseable remainder is left after the
loop, return the original file_content (rewinding a consumed seekable source) instead of
the truncated output. The multi-line happy path is unchanged.
* test(batches): mock resumable GCS upload in vertex batch prediction test
The vertex batch file-create path now streams to a GCS resumable session via
_aresumable_chunked_upload (httpx send) instead of AsyncHTTPHandler.post, so the
existing test's post mock no longer intercepted the upload and a real request hit
GCS (401). Mock _aresumable_chunked_upload to return the GCS object response; the
resumable protocol itself is covered in test_vertex_ai_files_streaming.py.
* fix(batches): resilient per-row token accounting; no hard-block on count failure
The batch input-file pass iterated a generator whose json.loads raised on a
malformed line; the outer except caught it and stopped the loop, so any body.model
on rows after a bad line was never collected and the model allowlist check ran
against a partial set. It also hard-blocked the batch with a 400 whenever token
counting raised, a backwards-incompatible change from the prior swallow-and-proceed
behavior that breaks legitimate rows the token counter cannot measure (e.g. some
multimodal content).
Iterate the JSONL line-by-line and account each row independently. A malformed line
is skipped (its request cannot run upstream anyway) and a row the counter cannot
measure falls back to a conservative size-based estimate. The loop never aborts, so
the allowlist check always sees every parseable model, and the token total is never
zeroed, so a crafted uncountable row still cannot evade the TPM limit, without
hard-rejecting a legitimate batch.
* perf(vertex/files): unblock async upload; drop empty finalize; widen batch MIME types
Three review follow-ups on the resumable batch upload:
- _aresumable_chunked_upload pulled chunks from a synchronous generator that runs
the per-row transform inline on the event loop thread, blocking other requests
between PUTs on large uploads. Each chunk is now produced via asyncio.to_thread.
- _iter_resumable_chunks no longer yields a trailing empty chunk, so an exactly
chunk-aligned upload finalizes on its last data chunk instead of an extra
zero-byte PUT; a 0-byte stream still finalizes via the caller's empty request.
- valid_content_type now accepts the MIME types clients label .jsonl batch uploads
with (text/plain, application/json, ndjson, ...), so such a batch file no longer
silently bypasses the streaming path into the buffered media upload.
* fix(vertex/files): keep legacy bucket_name as GCS bucket fallback
The rename to gcs_bucket_name dropped the legacy bucket_name key entirely, so an SDK caller passing bucket_name to a Vertex AI file create/retrieve/content call with GCS_BUCKET_NAME unset got ValueError("GCS bucket_name is required") where it previously resolved the bucket. _get_configured_bucket_name now reads gcs_bucket_name, then bucket_name, then the env var, and bucket_name is restored to OPTIONAL_KWARGS_KEYS so it survives get_litellm_params on the retrieve and content paths. gcs_bucket_name keeps precedence when both are present
* style: sort imports in llm_http_handler to satisfy I001 budget
---------
Co-authored-by: Yuneng Jiang <yuneng@berri.ai>
|
||
|
Sameer Kankute
|
e73cbfb026
|
fix(realtime): post-tool-call function_response id omission (#30446) | ||
|
Sameer Kankute
|
4c25b7a13d
|
chore: litellm oss staging (#30745)
* fix(proxy): bump health-check max_tokens default to 16 for GPT-5 compatibility (#30708) OpenAI GPT-5 models require max_completion_tokens >= 16. Health checks were using 5 (proxy/health_check.py) and 10 (health_check_helpers.py), causing failures on GPT-5 models. Fixes #23836 * fix: increase health check max_tokens from 5 to 16 (#23836) (#26610) GPT-5 models enforce a minimum of 16 for max_output_tokens. The current default of 5 still causes health checks to fail for these models. Bump the non-wildcard default to 16 — the smallest value that satisfies all known provider minimums while keeping health checks lightweight. Also tightens the wildcard test assertion from a weak disjunctive check to strict key-absence. Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix: ensure checks show gemini-3-flash-preview supports responseJsonS… (#30696) * fix: ensure checks show gemini-3-flash-preview supports responseJsonSchema. * fix: remove async keyword from test. * fix: make Bedrock Mantle Responses routing data-driven per model (#30700) * Make Bedrock Mantle Responses routing data-driven per model Route Bedrock Mantle models to the native Responses API based on each model's price-map capability signal instead of a hardcoded model-name heuristic, and derive the OpenAI-compatible base path segment per model. Responses dispatch now selects the native config when the model advertises responses support (/v1/responses in supported_endpoints, or mode=responses), both overridable via register_model and proxy model_info. This enables native Responses for gpt-oss-120b/20b and the gemma-4 family while keeping chat-only models (gpt-oss safeguard, nvidia, mistral, ...) on the existing chat-completions emulation. Capability is per-model, so gpt-oss-120b routes natively while gpt-oss-safeguard-120b does not despite sharing the gpt-oss substring. The wire path is a separate concern, driven by the existing use_openai_responses_path flag rather than a model-name match: gpt-5.x and gemma-4-* on /openai/v1, everything else (incl. gpt-oss) on /v1. The chat config now derives its base from the same flag, fixing gemma-4 chat-completions requests that previously went to /v1 instead of /openai/v1. Cost maps: add supported_endpoints to the gpt-oss entries (responses for the non-safeguard variants, chat-only for safeguard) and supported_endpoints + use_openai_responses_path to all three gemma-4 entries. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Address review: move capability helper into bedrock_mantle package Move the Responses capability check out of utils.py into litellm/llms/bedrock_mantle/common_utils.py as mantle_supports_responses, alongside its companion wire-path helper mantle_base_segment. Both are now pure functions of (model, model_cost): the price-map mode/supported_endpoints read replaces the get_model_info call, so the rules are unit-testable without patching global state and the Bedrock Mantle package is self-contained. Use str | None instead of Optional[str] on the new signatures to satisfy the ruff UP045 strict-rule gate. Add direct unit tests for both helpers. Fix test_register_model_restore_undoes_existing_key_overwrite: gpt-oss-120b now legitimately supports Responses, so it can no longer be the "None after restore" vehicle; use the chat-only safeguard variant, which isolates the register/restore effect from the model's own capability. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix(proxy): fail fast on non-PostgreSQL DATABASE_URL instead of hanging on startup (#30366) * fix(proxy): fail fast on non-PostgreSQL DATABASE_URL instead of hanging on startup LiteLLM's Prisma datasource is pinned to provider = 'postgresql', so a sqlite:// or mysql:// DATABASE_URL can never connect. Today that surfaces as an opaque startup stall where the port never binds, and a separate 'DB not connected' 500 on /key/generate when no DATABASE_URL is set at all leaves operators guessing what to configure. Validate the DATABASE_URL / DIRECT_URL scheme in run_server before any Prisma call and exit with an actionable message naming the unsupported scheme. Also reword CommonProxyErrors.db_not_connected_error to tell the operator to set DATABASE_URL to a postgresql:// connection string. Add regression tests covering postgres acceptance and sqlite/mysql/mssql rejection. * fix: resolve CI failures and proxy DB URL typing issue * fix(dashscope): treat an explicit 0.0 tier cost as a real price, not missing (#30653) The tiered cost calculator resolved a tier's per-token cost with `tier.get(cost_key) or tier.get(fallback_cost_key, 0)`. Because `or` short-circuits on any falsy value, a tier that legitimately prices a component at 0.0 (e.g. a free-cache-read tier with cache_read_input_token_cost: 0.0, or a free-reasoning tier) is treated as missing and silently billed at the full fallback rate (input_cost_per_token / output_cost_per_token). The flat-pricing path in the same module already handles this correctly with an `is None` guard. Resolve tier costs through a small helper that mirrors it, so 0.0 is honored at both the in-range and overflow sites. No shipped model currently has a 0.0 tier cost, so this is a latent defect; the fix makes the tiered path consistent with the flat path and prevents over-charging the first time such a tier appears. Adds unit tests covering the in-range and overflow paths, and drops an unused import flagged by ruff in the touched test file. * feat(proxy): show session-aggregate cost and duration in request logs (#25708) (#30507) * fix(anthropic): don't leak tool 'type' into OpenAI function parameters schema (#30618) In the messages->chat/completions bridge, translate_anthropic_tools_to_openai merged every non-mapped tool key into the function parameters dict. The Anthropic tool 'type' (e.g. 'custom') thus overwrote parameters.type ('object' -> 'custom'), and providers reject it ('custom' is not a valid JSON-Schema type). Exclude 'type' from the passthrough. Fixes #30557. * fix(proxy): stop IAM-refresh engine restart from cascading reconnects (#29176) (#30183) An RDS IAM token refresh recreates the Prisma client, which SIGKILLs the running query-engine and spawns a new one. That planned kill was indistinguishable from a crash, and three reconnect paths used two uncoordinated locks, so a single refresh triggered a cascade of engine kill/respawn cycles: 1. `_safe_refresh_token` (holds `_reconnection_lock`) -> recreate -> kill old engine, spawn new one. 2. The engine-death watcher sees that kill, assumes a crash, and calls `attempt_db_reconnect(force=True)` (a different lock, `_db_reconnect_lock`) -> recreate again -> kills the fresh engine. 3. In-flight queries failing during the swap are classified as transport errors and trigger their own `attempt_db_reconnect` -> recreate again. Fix coordinates planned restarts across the wrapper and the watcher: - PrismaWrapper records the old engine PID in `_expected_engine_deaths` before killing it; all four watcher death-detectors (waitpid thread, pidfd, already-dead probe, os.kill poll) consume that PID and skip the reconnect instead of treating it as a crash. - `recreate_prisma_client` now serializes through `_reconnection_lock` and bumps a monotonic `_engine_generation`. Callers pass `expected_generation` as an optimistic-lock token, so racing/cascading recreates collapse into a single restart (losers no-op). This closes the two-lock gap. - The direct reconnect path probes the writer with SELECT 1 before recreating; a healthy connection (e.g. engine already replaced by a refresh) skips the recreate entirely. - `_safe_refresh_token` coalesces: it skips when the current token still has more than the refresh buffer of runway, so stacked triggers (proactive loop + __getattr__ fallback) don't each restart the engine. An `on_engine_replaced` hook re-arms the watcher on the new PID. RoutingPrismaWrapper forwards `expected_generation` and skips recreating the reader when the writer recreate was skipped. * feat(bedrock): support file content retrieval for batch output files (#30595) Implements transform_file_content_request and transform_file_content_response in BedrockFilesConfig so GET /v1/files/{id}/content works for Bedrock batch files. The request transform resolves the file id (direct s3:// URI or base64 unified id) to its S3 object, validates bucket and key prefix against the server-configured bucket, and SigV4-signs an S3 GetObject using the same credential and region resolution as the existing upload path. The credential and region params are validated into a typed model at the boundary, so the only untyped values left are the botocore signing primitives. Also fixes the proxy managed-files path: CredentialLiteLLMParams now carries s3_bucket_name (previously dropped when building deployment credentials) and the managed-files hook passes the deployment credential snapshot when routing afile_content, so unified-id content retrieval works with per-model bucket config instead of only the AWS_S3_BUCKET_NAME env var. Preserves managed-file access control: the proxy file-content endpoint now rejects raw cloud-storage ids (s3://, gs://), which would otherwise skip the owner/team check that only runs for unified ids and let a caller read another tenant's batch output by its object key. Managed outputs are reachable only through their unified file id. The afile_content "not found" error now reports the caller's unified id rather than the resolved internal S3 URI. Fixes #16186, #15563 * fix(oci): make Cohere {{trace}} judges work (tool param types + agentic tool-calling continuation) (#30646) * fix(oci): map Cohere tool array/object params to lowercase builtins OCI's Cohere backend returns HTTP 500 on a tool parameter typed as a bare "List", which is what OCI_JSON_TO_PYTHON_TYPES produced for JSON-schema arrays. MLflow {{trace}} judges trip this: their tools (get_root_span, get_span) take an attributes_to_fetch array. The lowercase builtins list/dict are accepted; only the bare "List" 500s ("Dict" happens to be tolerated, but both are lowercased for consistency). Verified live against us-chicago-1 (cohere.command-a-03-2025 and command-latest). Adds a unit regression on the transformed parameterDefinitions plus a gated integration test exercising an array-param tool end to end. * fix(oci): make Cohere agentic tool-calling continuation work Two bugs broke the OCI Cohere tool-calling loop that MLflow {{trace}} judges drive once a tool has been executed and its result is fed back. Request side: litellm pulled the last user message into the top-level `message` and emitted the tool result as a TOOL entry in chatHistory. OCI rejects that ("cannot specify message if the last entry in chat history contains tool results"), and an empty message alone is rejected too ("message must be at least 1 token long or tool results must be specified"). OCI carries the current turn's results in a dedicated top-level `toolResults` field. The Cohere transform now sends an empty message, keeps the user turn in chatHistory, and puts the results in `toolResults`, matching the langchain-oracle reference. Tool results are no longer represented as chatHistory entries. Response side: tool-grounded answers come back with citations carrying `documentIds` (camelCase) and no `document_ids`, which made the required `CohereCitation.document_ids` field fail validation and sink the whole response parse. Those citations are never surfaced, so the field (and CohereSearchQuery's generation_id) is now optional. Verified live against us-chicago-1 (cohere.command-a-03-2025 and command-latest), single and multi-round tool loops. Adds unit regressions on the transformed request shape and on citation parsing, plus gated integration tests for the continuation. * feat: integrate Repelloai Argus guardrail (#30673) * feat(guardrails): add RepelloAI Argus guardrail integration (#1) * feat(guardrails): add RepelloAI Argus guardrail integration Add a new guardrail hook backed by RepelloAI Argus, with dashboard-managed asset policies enforced via an asset_id and X-API-Key auth. * fix(guardrails): harden RepelloAI Argus guardrail - scan streaming responses on output (was bypassing the guardrail) - log blocked verdicts as guardrail_intervened instead of success - treat auth/config errors (401/403/404/422) as misconfiguration that always blocks, not a fail-open-able unreachable error - default unreachable_fallback to fail_closed and read it directly; block on unknown/malformed verdicts so an API change can't silently disable enforcement - type unreachable_fallback as a Literal, drop the duplicate config model, expose unreachable_fallback in the config schema, and stop leaking the raw provider response / exception strings to the client * fix(guardrails): address RepelloAI Argus review feedback - support ARGUS_API_KEY (with REPELLOAI_API_KEY fallback) - make asset_id required in the config model - normalize unreachable_fallback so only fail_open opens; block on 400 misconfig - correct the shared unreachable_fallback field description * docs(guardrails): add RepelloAI Argus docs page and dashboard listing - add docs page covering config, env vars, modes, verdicts, failure semantics - list RepelloAI Argus in the Guardrail Garden with provider/logo mappings - add a regression test for the provider logo and display-name resolution * fix(guardrails): keep RepelloAI asset_id optional in config model A required asset_id leaked onto the shared LitellmParams (which inherits RepelloAIGuardrailConfigModel), breaking validation for every other guardrail. Keep it optional like sibling models; the guardrail __init__ still raises when asset_id is missing, which is the real enforcement. * Add comment for last user turn scanning * feat(guardrails): harden repelloai scanning * feat(guardrails): expand repelloai scanning to include tool definitions Add extraction of tool definitions and tool call arguments to the RepelloAI guardrail scanning. Improves detection coverage by including function schemas and parameters in the prompt sent to the guardrail service. Also captures detailed error responses in logs and adds guardrail header to streaming responses. * refactor(guardrails): fix and harden repelloai schema text extraction - Fix duplicate text in _iter_schema_text: previously all dict values were re-queued onto the stack even after scalar/list keys were already extracted explicitly, causing names/descriptions to appear twice in the scanned prompt - Extract schema key frozensets to module-level constants so they are not reconstructed on every call - Change _iter_schema_text from @classmethod to @staticmethod (cls unused) - Narrow _call_analyze stage param from str to Literal["prompt", "response"] - Add HttpxResponse type annotation to _raise_for_config_error - Add LLMResponseTypes annotation to async_post_call_success_hook response param * fix(guardrails): resolve pyright type errors in repelloai guardrail - Narrow async_handler.post return from Response|None to Response with explicit None guard before calling raise_for_status/json - Fix list comprehension returning str|None by switching to explicit loop with isinstance guard so pyright tracks the narrowing - Cast model_dump() result to Dict since hasattr does not narrow object type in pyright * fix(guardrails/repello): include Responses API instructions field in prompt scan The /v1/responses top-level `instructions` field was not included in _extract_prompt_text, allowing a caller to bypass guardrail policy checks by putting blocked content in `instructions` while keeping `input` benign. * feat: add api_key to config model and read prompt from data dict * fix(guardrails/repello): plug input_text and tool-call response bypass gaps Responses API input content parts with type 'input_text' were silently dropped by build_inspection_messages (which only handles type='text'), allowing callers to send blocked content via that path without triggering the pre-call scan. Fix: add _extract_input_text_parts to RepelloAIGuardrail and call it when walking the Responses API input messages. Post-call scanning skipped responses whose choices contained only tool_calls or function_call (message.content=None), letting models put blocked output in function arguments undetected. Fix: _extract_chat_completion_text now calls _extract_tool_call_args_from_message on each choice message. Also replace typing.Dict/List with builtin dict/list to clear TID251 strict ruff violations introduced by this file. * fix(guardrails/repello): scan Responses API function_call output arguments Output items with type 'function_call' in a /v1/responses response were skipped by _extract_responses_api_text; only 'message' items were walked. A model could return blocked content in function_call.arguments undetected. Now extract arguments from function_call output items before scanning. * refactor(guardrails/repello): clean up typing and remove lint-any workarounds - Replace Optional[X]/Union[X,Y] with X|None/X|Y union syntax throughout - Use dict[str, object] instead of bare dict in all signatures - Remove **kwargs from __init__; declare guardrail_name, event_hook, default_on explicitly - Replace getattr(litellm_params, ...) with direct attribute access now that LitellmParams inherits RepelloAIGuardrailConfigModel - Add _event_hook_from_mode() to convert str|list[str]|Mode to typed GuardrailEventHooks - Use TypeAdapter.validate_json() instead of response.json() + manual dict construction - Add _is_object_dict/_is_object_list TypeGuard helpers to narrow object types without Any - Remove cast() workarounds and typed intermediate variables that existed only for the now-removed lint-any CI check - Drop _AddLiteLLMCallback Protocol; budget has sufficient slack for the one reportUnknownMemberType - Fix GuardrailConfigModel missing type arg: GuardrailConfigModel[BaseModel] * fix(guardrails/repello): suppress LIT007 on TypeGuard helpers and add streaming scan-skip warning - Add guard-ok suppressions to _is_object_dict and _is_object_list to satisfy the LIT007 hard-zero budget gate - Emit verbose_proxy_logger.warning when the streaming hook finds no inspectable text after assembly, matching observability of pre/post hooks * refactor: modifications for lint check * feat: add Pinstripes as an OpenAI-compatible provider (#30567) * feat: add Pinstripes as an OpenAI-compatible provider Pinstripes (https://pinstripes.io) is an OpenAI-compatible inference provider serving open-source models (GLM-4.5-Air, Qwen3, DeepSeek, etc.) with per-token pricing and no subscriptions. Changes: - `litellm/llms/openai_like/providers.json`: register pinstripes with base_url, api_key_env, and max_completion_tokens→max_tokens mapping - `litellm/types/utils.py`: add `PINSTRIPES = "pinstripes"` to LlmProviders - `litellm/constants.py`: add to openai_compatible_providers and openai_compatible_endpoints lists - `litellm/litellm_core_utils/get_llm_provider_logic.py`: auto-detect provider when api_base is "https://pinstripes.io/v1" - `provider_endpoints_support.json`: document supported endpoints - `tests/`: 7 unit tests covering provider registration, resolution, URL auto-detection, api_base override, and Router config Usage: import litellm response = litellm.completion( model="pinstripes/ps/glm-4.5-air", messages=[{"role": "user", "content": "Hello"}], api_key=os.environ["PINSTRIPES_API_KEY"], ) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(pinstripes): resolve Greptile P1 review comments - Add api_base_env: PINSTRIPES_API_BASE to providers.json so env var override works - Set responses: false in provider_endpoints_support.json — not actually wired up - Remove docs/my-website/docs/providers/pinstripes.md — belongs in litellm-docs repo Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(pinstripes): add api_base_env and correct responses capability - Add api_base_env: PINSTRIPES_API_BASE to providers.json - Set responses: false in provider_endpoints_support.json Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(pinstripes): wire up Responses API — add supported_endpoints Adds supported_endpoints: ["/v1/chat/completions", "/v1/responses"] so JSONProviderRegistry.supports_responses_api returns true correctly, matching what provider_endpoints_support.json advertises. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(pinstripes): enable embeddings endpoint Pinstripes serves nomic-embed-text-v1.5 and bge-m3 via /v1/embeddings. Add /v1/embeddings to supported_endpoints and set embeddings: true. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(pinstripes): use 4-space indentation in model_prices_and_context_window.json Matches the file's existing convention. Flagged by Greptile review. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(pinstripes): set a2a: false — A2A protocol not implemented All comparable JSON-configured providers (tensormesh, parasail, empiriolabs, libertai, neosantara) have a2a: false. Pinstripes does not implement the Google A2A protocol, so this should be false to match. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: inference_provider <max@redactedlab.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(rag): attach existing OpenAI file ids (#30628) * fix(rag): attach existing OpenAI file ids * chore: use modern typing in rag ingest fix * chore: retrigger ci * fix(anthropic-messages): apply cache_control_injection_points on /v1/messages path (#30341) cache_control_injection_points was only consumed by the chat/completions prompt-management hook; on the native Anthropic /v1/messages path it was forwarded unused, so deployment-level cache injection was silently dropped (cache_creation_input_tokens stayed 0 for Anthropic-native clients). Add AnthropicCacheControlHook.apply_to_anthropic_messages_request to inject cache_control at block level for system / tools / message locations (the only forms /v1/messages accepts), wire it into the native anthropic_messages handler, and pop the param so it does not leak upstream as an unknown field. A {location: message, role: system} config is redirected to the top-level system prompt so the same YAML works on both endpoints. Injection respects Anthropic's 4-block cache_control limit shared across system, tools, and messages: client-supplied markers count toward the cap and are never overwritten, a slot is reserved per Bedrock tool_config point, and injection stops once the budget is exhausted. Locations this path cannot represent (tool_config) are forwarded downstream instead of being silently consumed, mirroring get_chat_completion_prompt's remaining_points pass-through. Built on litellm_internal_staging. Refs BerriAI/litellm#30293 * fix(proxy): release budget reservation when a request is cancelled mid-flight (#30522) * fix(proxy): release budget reservation on cancel when no chunk was delivered The pre-call budget reservation increments the cross-pod spend counter by a request's worst-case cost, then reconciles it on success (cost callback) or error (failure hook). A client disconnect or timeout cancels the request and surfaces as CancelledError / GeneratorExit, which neither path catches, so the reservation leaks. Under a retry storm the leaked holds accumulate, pin the counter above real spend, and return spurious 429 "Budget has been exceeded" to keys whose spend is far below budget; the counter only recovers when its TTL lapses, so the failure is intermittent and self-healing. Release the reservation in async_streaming_data_generator (which the Anthropic and Google SSE generators delegate to) on the (CancelledError, GeneratorExit) path, alongside the existing max_parallel_requests release. release_budget_ reservation_on_cancel runs under asyncio.shield so it completes despite the in-progress cancellation, is guarded by the reservation's finalized flag, and swallows a failing release so it cannot replace the in-flight cancellation. The refund is gated on whether a chunk reached the client. The flag is set immediately before the yield, after the slow-path hook await: an async generator suspends at the yield, so a GeneratorExit on disconnect after a delivered chunk sees it True (keep the hold), while a cancellation during the slow-path await leaves it False (refund, nothing sent). A non-streaming cancellation delivers nothing and a completed non-streaming response is reconciled by the success callback, so neither needs a release here. Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(proxy): reconcile a cancelled reservation to input cost, not zero A streaming request cancelled before the first chunk previously reconciled its reservation to zero and finalized it. But by the time the generator is consuming the response the provider call was already dispatched, so the input tokens were billed even though no chunk reached the client, and the success/failure cost callbacks are skipped on cancellation. Refunding to zero let a caller send an expensive request and abort pre-token to dodge the input charge. Compute the request's input-token cost at reservation time and reconcile the cancelled reservation to it instead of zero. The worst-case output portion of the reservation is still released (so a legitimate mid-flight cancellation no longer pins the counter and 429s the key), while the input the provider already processed is charged. --------- Co-authored-by: Bytechoreographer <Bytechoreographer@users.noreply.github.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(caching): encode object name in GCS cache GET path (#30378) GCS cache reads always missed when gcs_path was set. The GET methods interpolated the object name directly into the URL path, while the GCS JSON API requires it to be URL-encoded (a "/" must be sent as %2F). With gcs_path configured the object name is "<prefix>/<sha256>", so the raw slash produced a malformed object path and GCS returned 404. httpx does not raise on 4xx, so the status_code == 200 check fell through and get/async_get returned None, silently missing on every read. Without gcs_path the key has no slash, which is why this went unnoticed. Wrap the object name with urllib.parse.quote(..., safe="") in get_cache and async_get_cache. Apply the same encoding to the name= query parameter in set_cache and async_set_cache so the key written matches the key read back. Adds regression tests asserting the GET path and SET query are encoded (%2F) when gcs_path is set, for both sync and async paths; these fail on the unpatched code. Fixes #30377 * chore: add soniox stt-async-v5 model (#30672) * fix(proxy): include model group aliases in v1 model info (#30626) * Include model group aliases in v1 model info * Fix model info alias implementation * removed extra blank line * chore: rerun CI * fix(lint): remove redundant noqa directive in proxy_cli.py * fix: address greptile review - restore bedrock_mantle auth symbols, guard OCI empty message list, validate DIRECT_URL scheme * Revert "fix: address greptile review - restore bedrock_mantle auth symbols, guard OCI empty message list, validate DIRECT_URL scheme" This reverts commit |
||
|
Sameer Kankute
|
1ccc1e5b23
|
chore: litellm oss staging160626 (#30527)
* feat(ui): gate "Default Credentials" hint on /ui/login behind env flag (#30234) Adds LITELLM_HIDE_DEFAULT_CREDENTIALS_HINT (and an equivalent general_settings.hide_default_credentials_hint) that suppresses the "By default, Username is admin and Password is your set LiteLLM Proxy MASTER_KEY" info card rendered on /ui/login and /fallback/login. Motivation: in production deployments operators set UI_USERNAME / UI_PASSWORD (or SSO), and the hardcoded hint becomes factually incorrect and is flagged by security scanners (Tenable WAS plugin 114625) as information disclosure. There is currently no way to suppress it without forking the dashboard. Behaviour: - Default is unchanged (hint shown), so existing deployments are unaffected. - New field hide_default_credentials_hint on the well-known UI config endpoint, populated from the env var or general_settings. - LoginPage.tsx conditionally renders the Alert based on the flag. Refs: BerriAI/litellm#30232 * fix(router): clean pattern_router state on upsert/delete (#29601) * fix(router): clean pattern_router state on upsert/delete PatternMatchRouter.add_pattern was append-only, and neither Router.upsert_deployment nor Router.delete_deployment removed the existing entry. Rotated-out api_keys stayed in the routing rotation for wildcard deployments (model_name with `*`) until proxy restart, silently defeating key rotation as an admin operation. The same leak applied to provider_default_deployment_ids and per-team pattern routers, and the patterns list grew unboundedly on every edit * test(router): direct unit tests for _remove_deployment_from_wildcard_state router_code_coverage.py greps test files for AST Call nodes and flagged the helper as untested because the existing coverage only exercised it transitively through upsert/delete. Adds two direct tests that pin the helper's contract (cleans across global pattern router, per-team routers with empty-router pop, and provider_default_deployment_ids; noop on falsy model_id) * fix(router): address Greptile review on pattern_router cleanup Widen PatternMatchRouter.remove_deployment annotation to Optional[str]; the implementation already handles None via the falsy guard and the unit test exercises it directly. Move _remove_deployment_from_wildcard_state up one level in upsert_deployment so it runs whenever the prior deployment is on the router, not only when the model_id is present in the fast-mapping index. The scenario is currently unreachable (get_deployment shares the same index), but the cleanup is idempotent so this is defensive against any future divergence between those code paths. * fix(router): widen _remove_deployment_from_wildcard_state to Optional[str] Moving the call out of the inner `deployment_id in deployment_fast_mapping` block in the previous commit lost mypy's narrowing of `deployment_id` from Optional[str] to str, tripping the lint CI. The helper already handles None via its falsy guard, so widening the annotation matches the actual contract. * fix(router): make delete_deployment wildcard cleanup symmetric with upsert After the previous commit moved _remove_deployment_from_wildcard_state out of the inner index-map guard in upsert_deployment, delete_deployment was still calling it only inside `if deployment_idx is not None`. Greptile flagged the asymmetry: under a desynced index_map, delete would silently leave the stale wildcard credential in pattern_router. Moves the cleanup call to the top of the try block, mirroring the upsert path. Cleanup is idempotent so the change is a no-op on the happy path. Adds a regression test that simulates the desync by removing the entry from model_id_to_deployment_index_map and asserts delete still clears pattern_router. * fix(pricing): add 1h cache-write cost for Anthropic Sonnet 4.5/4.6 (#30474) The native anthropic claude-sonnet-4-5/4-6 price-map entries were missing cache_creation_input_token_cost_above_1hr (and the >200K long-context sub-tier for 4.5), so 1-hour-TTL cache writes were costed at the 5-minute rate. Adds 6e-06 regular (and 1.2e-05 long-context) = 2x base input, matching the vertex_ai/azure_ai/bedrock siblings and the older claude-sonnet-4-20250514 entry. Adds a regression test. * fix(proxy): cancel upstream gemini request and release httpx connection on client disconnect (#30075) * fix(proxy): cancel upstream gemini request and release httpx connection on client disconnect - add _check_request_disconnection to common_request_processing; wrap llm_call as asyncio.Task so it can be cancelled; catch CancelledError and raise HTTPException(499) when client disconnects before LLM responds (non-streaming path) - pass raw httpx.Response into ModelResponseIterator in make_call/make_sync_call so the iterator holds a reference to the underlying connection - implement ModelResponseIterator.aclose() and .close(): close the line iterator then explicitly call response.aclose()/response.close() to release the httpx connection when the client drops mid-stream; errors are debug-logged, not raised - add tests for _check_request_disconnection (cancels task, graceful on exception, does not cancel when client stays connected) and base_process_llm_request 499 behavior; add TestModelResponseIteratorCleanup verifying aclose/close propagation through CustomStreamWrapper * fix(proxy): record 499 on streaming disconnect and cancel orphaned gather tasks Wire streaming generator cleanup to log client_disconnected with error_code 499 in spend logs, cancel pending during_call_hook tasks when the LLM call is cancelled on disconnect, and align the 600s poll limit comment with proxy_server. * fix: extract client disconnect logging helper to satisfy PLR0915 * fix: resolve mypy and code-quality CI failures for client disconnect logging Cast client disconnect error_information for mypy, only await pending gather tasks to avoid masking LLM errors, and add tests for the new logging helper and gather cleanup. * fix(proxy): harden gather cleanup so finally cannot mask LLM errors * fix(proxy): shield streaming disconnect logging and strip spoofable metadata Move streaming disconnect recording into a shielded cancel scope, add gather cleanup regression coverage for guardrail-converted cancels, and strip client_disconnected/error_information from user metadata at the proxy boundary. * fix(proxy): only map CancelledError to 499 for client disconnect Track when the disconnect poller cancels the LLM task and re-raise other CancelledError paths so graceful shutdown is not reported as HTTP 499. * fix(proxy): remove dead _check_request_disconnection helper Non-streaming client disconnect is handled by staging's cancel_on_disconnect path via _await_llm_call_cancelling_on_disconnect. Drop the unused is_disconnected poller and its unit tests; rename the remaining integration tests to TestDisconnectGatherCleanup. * feat(mistral): add mistral-medium-3-5 to model_prices_and_context_wind.. (#29303) * feat(mistral): add mistral-medium-3-5 to model_prices_and_context_window.json Mistral's docs page lists mistral-medium-3-5 as a new model offering. Pricing/specs sourced from Mistral's published model metadata: - input: $1.50 / 1M tokens - output: $7.50 / 1M tokens - context: 262,144 tokens - capabilities: vision, function calling, structured outputs, assistant prefill Adds entry: `mistral/mistral-medium-3-5`, mirroring the pattern used for the rest of the Mistral family. test(mistral): add model_info test for mistral-medium-3-5 + sync backup cost map - Mirror mistral/mistral-medium-3-5 entries into litellm/model_prices_and_context_window_backup.json so the bundled model cost map matches the canonical model_prices_and_context_window.json. - Add tests/test_litellm/test_mistral_medium_3_5_model_metadata.py covering pricing tiers, capability flags, context window, provider routing, and parity between the main and backup cost maps. - Point 'source' at the live Mistral models documentation page. * fix(ui): three small UI fixes — Gemini api_base + credential form reset + Mode badge (#30419) * fix(ui): three small UI fixes — Gemini api_base field + credential form reset + Mode badge Three independent fixes; bundled because they all touch the credential-form / logging-callbacks area. 1. expose api_base field on Google AI Studio credential form The runtime gemini provider supports custom api_base via `vertex_llm_base._check_custom_proxy`; the UI just needs to expose the field. Adds api_base to the Google_AI_Studio credential form ordered before api_key (matching OpenAI/Anthropic conventions). Default value matches the canonical Google AI Studio endpoint that LiteLLM's gemini provider talks to when api_base is unset, so leaving the default in the form behaves identically to leaving it blank. 2. reset credential form state when switching providers Switching the Provider select in AddCredentialModal / EditCredentialModal left the previous provider's field values populated. The form then submitted a mixed payload (e.g. Azure deployment fields under an OpenAI credential), producing confusing failures. Extract `getProviderFieldDefaults` helper and reset the form to it on provider change. Unit-tested via the extracted helper because Antd Select's portal/dropdown behaviour is unreliable in jsdom. 3. logging callbacks table reads backend `type` for Mode badge (#35) The `/get_callbacks` proxy endpoint returns each callback as `{name, type, variables}` where `type` is `"success"` or `"failure"`. The same callback name can appear twice (one per event class) and the two entries fire on disjoint events. `LoggingCallbacksTable` ignored `type` and read `record.mode` (always undefined), so every row fell back to the "Success" badge. A `generic_api` callback registered for both classes showed up as two identical "Success" rows + React duplicate-key warning. Read `record.type` first (fall back to `record.mode` for newly- added not-yet-server-acknowledged rows). Composite rowKey `${name}-${type ?? mode ?? 'success'}`. Removed leftover debug `console.log`. * fix(ui): drop api_base default_value to preserve Gemini v1alpha auto-routing Greptile P2 (PR #30419, threads on lines 1255-1256 of provider_create_fields.json): the api_base field's `default_value` was hard-coded to "https://generativelanguage.googleapis.com/v1beta". This: 1. Bakes v1beta into every credential record saved through the form, even when the user never touched the field. If LiteLLM's internal gemini default URL ever changes, those persisted credentials keep hitting the stale path. 2. Bypasses `_get_gemini_url`'s automatic version routing for Gemini 3+ models. That helper picks v1alpha for Gemini 3+ and v1beta for older models when api_base is unset. With the default pre-filled (and `_check_custom_proxy` then taking over because api_base is non-empty), Gemini 3+ requests get pinned to v1beta and may fail or behave unexpectedly — purely because the user accepted the visible default. Fix: set `default_value` to `null` and move the canonical URL guidance into the `placeholder` (visible to the user, never persisted) and an expanded tooltip. UX is unchanged — the URL is still shown in the greyed-out input — but the auto-version-routing path stays default. Updated test_google_ai_studio_provider_fields_expose_api_base to assert the new contract (`default_value is None`, `placeholder` carries the canonical URL), with a comment pointing at the Greptile threads as the rationale so future contributors don't accidentally re-introduce the default. 26/26 tests in the file pass. JSON validates (`json.load` clean). * feat(azure_ai): add gpt-5.5 to model cost map (#30428) * feat(azure_ai): add gpt-5.5 to model cost map Adds azure_ai/gpt-5.5 and its dated snapshot azure_ai/gpt-5.5-2026-04-23 to both the canonical and bundled cost maps. gpt-5.5 is generally available on Azure AI Foundry; pricing mirrors the openai gpt-5.5 entry, matching the established azure_ai convention (verified identical for gpt-5.4), in the azure tier structure (base / above-272k / priority). supports_minimal_ reasoning_effort is false, the capability that changed from gpt-5.4. Fixes #30306 * Update tests/test_litellm/test_gpt_5_5_model_metadata.py Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix: guard check_and_fix_namespace against None key (#30435) * fix: guard check_and_fix_namespace against None key When user_id is None, the cache key can be None, causing AttributeError: 'NoneType' object has no attribute 'startswith' in check_and_fix_namespace. Add an early return for None key to prevent the error and the ERROR-level log noise it produces on every unauthenticated request. Fixes #30424 * fix: update type annotations for check_and_fix_namespace - key: str -> Optional[str] (now handles None input) - return: str -> Optional[str] (returns None when input is None) Addresses Greptile review concern about type signature mismatch. * fix: revert check_and_fix_namespace type signature to str to fix MyPy downstream errors * fix: update type annotations for check_and_fix_namespace - Change signature from str -> str to Optional[str] -> Optional[str] - Remove type: ignore comment on None return - Add None guard in async_set_cache_sadd before passing to helper Addresses review feedback from Sameerlite on type mismatch. * Revert "fix: update type annotations for check_and_fix_namespace" This reverts commit |
||
|
Sameer Kankute
|
816fca939f
|
chore(oss): litellm oss staging 150626 (#30463)
* fix(pricing): add GitHub Copilot MAI Code Flash pricing (#30415) * fix(pricing): add GitHub Copilot MAI Code Flash pricing Add GitHub Copilot pricing entries for MAI-Code-1-Flash and the internal Copilot CLI model name so cost calculation can price input, cached input, and output tokens. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * test(pricing): cover GitHub Copilot MAI Code Flash pricing Add regression coverage for both GitHub Copilot MAI-Code-1-Flash model names, including cached input pricing, chat endpoint metadata, and cost_per_token arithmetic. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * fix(router/proxy): propagate completed_response through FallbackResponsesStreamWrapper for streaming /v1/responses container ownership (#30210) (#30213) * fix(router/proxy): propagate completed_response through FallbackResponsesStreamWrapper for streaming /v1/responses container ownership (#30210) #28990 added ownership recording for streaming /v1/responses via _wrap_responses_stream_for_container_ownership, which reads `getattr(stream_response, 'completed_response', None)` to extract the ResponsesAPIResponse. The unit test bypassed the Router, so it never exercised the production wrapping path. Through the Router (every proxy deployment), the stream is wrapped by FallbackResponsesStreamWrapper (router.py:2527). Its __init__ set `self.completed_response = None` and __anext__ only forwarded chunks — the inner source iterator's terminal event never bubbled up to the attribute the ownership hook reads, so the hook silently recorded nothing and every follow-up /v1/containers/<id>/files call returned 403 for non-admin keys. This commit: - router.py: pre-resolves the responses-API terminal event tuple (response.completed / .incomplete / .failed) once per _aresponses_streaming_iterator call, and has the wrapper's __anext__ sniff each forwarded chunk's .type. First terminal event hit gets stored on the wrapper's completed_response. Iterator-agnostic — works for source_iterator AND any future wrapper. - common_request_processing.py: when _extract_completed_responses_response returns None we now warn instead of silently skipping. Reporter on #30210 lost a day to this exact silent skip; the warning surfaces future regressions of the same shape directly in operator logs. Fixes #30210 * fix(router): type-ignore wrapper getattr-defaults; broaden ownership-skip warning CI lint (mypy) flagged the three pre-existing getattr(..., None) assignments in FallbackResponsesStreamWrapper.__init__: router.py:2564 self.response = getattr(source_iterator, 'response', None) router.py:2565 self.model = getattr(source_iterator, 'model', None) router.py:2566 self.logging_obj = getattr(..., None) Those lines also exist on litellm_internal_staging and pass mypy there. Adding the typed terminal-event tuple above the class made the function body more narrowable, which surfaced the pre-existing mismatch — base class declares non-Optional types but the bridge path (LiteLLMCompletionStreamingIterator) legitimately omits these. Keep the None fallback and silence with type: ignore[assignment]. Greptile 4/5 note: the ownership-skip warning hard-named code_interpreter which misleads operators when a non-code_interpreter stream aborts. Generalize to 'any tool container (e.g. code_interpreter)'. * fix(register_model): drop synthesized zero costs to preserve sparse entries (#30198) (#30201) * fix(register_model): drop synthesized zero costs to preserve sparse entries (#30198) get_model_info synthesizes input_cost_per_token / output_cost_per_token = 0 when they are absent from the raw entry (the price-unknown and free cases share the same representation). register_model then merges that result back into litellm.model_cost, which flips a sparse entry from 'no cost keys' (priced via model name) to 'cost keys = 0' (free). That defeats _is_cost_explicitly_configured (#24949) on re-registration: _is_model_cost_zero returns True, common_checks skips every tag / key / team / user / org budget check for the group, and over-budget traffic keeps returning 200. Spend keeps recording because cost calc still resolves by model name, so the symptom is silent and only triggers on the second register_model pass (router rebuild, /model/update, config sync). Mirror the existing litellm_provider-None guard one block above and pop the cost fields from the synthesized result when they are absent from the raw entry and not in the caller's value. Caller-provided zeros (genuinely free models, BYOK overrides) are preserved. Fixes #30198 * fix(register_model): switch _raw_entry to is-None checks + drop dead test assertion Greptile #30201 review notes: - the `or`-chain in the raw-entry lookup treated an empty dict (a key with no fields) as falsy and fell through to the second arm — replace with explicit `is None` checks so a present-but-empty entry is still taken at face value. - the first assertion in `test_router_double_init_keeps_db_model_entry_sparse` used `in (None, 0)` which passes under the bug condition (cost = 0 matches the tuple); the strong follow-up assertion already covers every shape, so drop the dead branch. * fix(bedrock mantle): use unique function-call id for responses->chat tool calls (#30426) * fix(bedrock mantle): use unique function-call id for responses->chat tool calls ... * fix(bedrock mantle): scope unique tool-call id fallback to degenerate call_id The previous revision preferred the Responses item id for every tool call, which broke providers (and existing tests) where call_id is a unique, canonical correlation key. Restrict the fallback to the degenerate index-based call_id that Bedrock Mantle returns (call_0, call_1, ... resetting per response) and keep call_id otherwise. Revert the change to the OUTPUT_ITEM_DONE streaming handler, whose tool_call_chunk is never emitted (dead code, per review). Extend the regression tests to assert a normal call_id is preserved. * fix(router): preserve azure_ad_token through CredentialLiteLLMParams for /v1/files + batches (#30235) (#30241) * fix(router): preserve azure_ad_token through CredentialLiteLLMParams for /v1/files + batches (#30235) Router.get_deployment_credentials_with_provider re-validates a deployment's litellm_params through CredentialLiteLLMParams before handing them to file/batch/passthrough callers: return CredentialLiteLLMParams( **deployment.litellm_params.model_dump(exclude_none=True) ).model_dump(exclude_none=True) Any field NOT declared on CredentialLiteLLMParams gets silently dropped on the way through. azure_ad_token was undeclared, so Azure deployments using OAuth/M2M (azure_ad_token instead of a static api_key) silently lost their token at the files endpoint and the proxy returned: Missing credentials. Please pass one of api_key, azure_ad_token, azure_ad_token_provider, ... Declare azure_ad_token on CredentialLiteLLMParams alongside api_key / api_base / api_version so it rides through the round-trip. Static-key deployments stay unaffected (Optional, default None, dropped by exclude_none=True). Provider-callable (azure_ad_token_provider) is a separate concern and out of scope here. Fixes #30235 * fix(ui-types): regenerate schema.d.ts for new azure_ad_token field CI's 'Verify schema.d.ts matches the proxy OpenAPI spec' check auto-detected the new field and emitted the exact diff to apply. Two schemas had `aws_secret_access_key` from CredentialLiteLLMParams, both get the new azure_ad_token marker next to it. * fix(proxy): org_admin with own user_id now sees all org teams on /v2/team/list (#30247) When the UI sends the callers own user_id (as it does for non-Admin global roles), _enforce_list_team_v2_access now nulls it out for org admins so _build_team_list_where_conditions scopes by organization_id only -- matching the legacy /team/list behavior and the documented intent. Fixes #30215 Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * test(vertex_ai): multi-region regression coverage for cachedContents host (#29571) (#29707) litellm_internal_staging already routes the cachedContents URL through get_vertex_base_url, fixing the multi-region 404 reported in #29571 — but carries no test coverage for the actual regression scenario (eu/us must resolve to the REP host aiplatform.{geo}.rep.googleapis.com). Add TestContextCachingMultiRegionUrls: parametrized eu/us REP-host assertions (including absence of the old broken {geo}-aiplatform host), plus regional (us-central1) and global no-regression checks. * fix(proxy): close upstream LLM stream when client disconnects mid-stream (#30245) * fix(proxy): close upstream LLM stream when client disconnects mid-stream When a streaming client disconnects, Starlette abandons the response body iterator without calling aclose(), so the proxy's connection to the upstream backend stays open until garbage collection, which may never come. The backend (e.g. vLLM) keeps generating into a dead pipe: small responses drain invisibly into TCP buffers while large ones block the backend on a full send buffer indefinitely (observed via lsof as an ESTABLISHED proxy->backend connection minutes after the client left) create_response now returns a StreamingResponse subclass that closes both its body iterator and the wrapped upstream-facing generator in a shielded finally. The upstream generator is closed directly rather than through a cascade because aclose() on a never-started generator skips its body, which would make the cascade a no-op when the client disconnects before the first chunk is sent. async_streaming_data_generator also gains the same shielded finally-aclose that async_data_generator in proxy_server.py already had, covering the Anthropic and Google SSE paths With this, killing a streaming client causes the backend to observe the abort within about a second and free its slot, while completed streams are unaffected. No flag is needed, unlike the non-streaming opt-in cancel in #30223: this only releases resources after the client is already gone and does not change any response a client can observe Fixes #30244 * fix(proxy): close upstream even when body iterator aclose raises BaseException Addresses the Greptile finding on #30245: the cleanup loop caught only Exception while the generator-level cleanup catches BaseException, so a CancelledError or GeneratorExit escaping body_iterator.aclose() would skip closing the upstream generator. Both sites now use the same scope and a regression test pins that the upstream is closed even when the body iterator explodes with a BaseException * fix(llms): expose aclose on BaseModelResponseIterator so stream close reaches the provider connection The response-level close added for #30244 only worked for SDK-based providers (e.g. openai), whose streams expose aclose all the way down. Providers served by base_llm_http_handler (hosted_vllm and most modern transformation-based providers) wrap a bare response.aiter_lines() generator in BaseModelResponseIterator, which had no aclose or close at all, and nothing retained the httpx response object; so CustomStreamWrapper.aclose() silently did nothing and the upstream connection stayed open. Verified with a vLLM-style mock: with hosted_vllm/ the backend streamed all 100 chunks to completion after the client disconnected, while openai/ aborted at chunk 6 BaseModelResponseIterator now carries an optional http_response and an aclose() that closes it; make_async_call_stream_helper attaches the response after building the iterator. With this, hosted_vllm aborts the backend within ~1.6s of the client dropping, and completed streams are unaffected --------- Co-authored-by: kursad <kursad.lacin@brado.net> * feat(anthropic): surface compaction usage iterations data (#27065) * feat(anthropic): surface compaction usage iterations data * style: apply black formatting to fix lint checks * fix(usage): correct calculate usage with cached tokens when use ChatCompletionUsageBlock (#30422) * fix(usage): correct calculate usage with cached tokens when use ChatCompletionUsageBlock * fix(usage): optimize test imports * feat: add fastCRW search provider (#30434) * feat(provider): add LibertAI as a JSON-configured OpenAI-compatible provider (#30203) * feat(provider): add LibertAI as a JSON-configured OpenAI-compatible provider * libertai: update served endpoints backup + add mode/matrix tests Addresses review feedback: - Add libertai to litellm/provider_endpoints_support_backup.json, the file actually served by GET /public/supported_endpoints (the root provider_endpoints_support.json already had it). - Add tests asserting bge-m3 normalizes to mode='embedding' and that the served matrix lists libertai. embeddings stays false: the JSON-configured provider path only wires chat routing (OpenAILike embedding handler is reached only for literal openai_like/llamafile/lm_studio), matching the llamagate precedent; bge-m3 remains in the cost map for metadata. --------- Co-authored-by: Moshe Malawach <moshemalawach@users.noreply.github.com> * feat(provider): add ModelScope as an OpenAI-compatible provider (#28460) * add ModelScope API support * add modelscope api support * update modelscope model list * add image-genetation support * update test and multimodal * fix: address PR review feedback for modelscope provider * update README * fix(customer_endpoints): restrict /customer/daily/activity to admin-only (#28849) * fix(customer_endpoints): restrict /customer/daily/activity to admin-only * fix(customer_endpoints): check role before prisma_client guard * fix(custom_guardrail): key disable_global_guardrails takes precedence over team guardrail list (#28563) * fix(fallbacks): preserve fallback model in SDK fallback responses (#28260) * fix(fallbacks): preserve fallback model in response when using SDK-level fallbacks * fix(fallbacks): gate x-litellm-* passthrough to trusted callers only The previous patch unconditionally let `x-litellm-*` keys bypass the `llm_provider-` prefix in `process_response_headers`. That function is also called on raw upstream-provider response headers (e.g. from `llm_http_handler.py`), so a malicious provider could return `x-litellm-attempted-fallbacks` and spoof a LiteLLM-internal marker, bypassing the proxy model-override guard. Add a `preserve_litellm_internal_headers` flag (default False). Only `response_metadata.py`, which re-processes the already-built `_hidden_params["additional_headers"]` dict (LiteLLM-owned), passes True. Raw provider header callsites keep the default False, so upstream `x-litellm-*` still gets the `llm_provider-` prefix. Adds a regression test for the spoofing case and renames the existing preserve test to make the trusted-path semantics explicit. * fix(fallbacks): ignore preserve_litellm_internal_headers for raw httpx.Headers inputs * style(core_helpers): apply black formatting * fix(lint): remove banned typing.List/Dict/Any imports and suppress PLR0913 on interface overrides Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(lint): apply black formatting to modelscope chat transformation Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(lint): replace noqa with proper fixes — use **kwargs and Awaitable instead of Any/List Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(lint): remove unused AllMessageValues import Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * revert: restore base_model_iterator.py to original PR state Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(lint): restore full method signatures for MyPy compatibility; bump PLR0913 budget for new provider files Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(lint): use @override to suppress PLR0913 on inherited signatures instead of bumping budget The overrides keep their full base-class signatures for MyPy compatibility, but those signatures carry more than five parameters, which tripped PLR0913 on each subclass redeclaration. Since the arity is dictated by the base class and cannot be reduced, decorate the overrides with typing_extensions.override; ruff treats that as the intended signal that the parameter count is not under the author's control and skips PLR0913. This restores the PLR0913 baseline to 1813. * fix(lint): add @override to modelscope image generation overrides Apply the same typing_extensions.override treatment to the image generation config so its inherited-signature overrides do not count against PLR0913. --------- Co-authored-by: Joel Tony <github@jaytau.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: hcl <chenglunhu@gmail.com> Co-authored-by: ztko <96878659+koztkozt@users.noreply.github.com> Co-authored-by: Nahrin <nahrin@nahrinoda.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> Co-authored-by: Humphrey <a739376838@gmail.com> Co-authored-by: kursadlacin <kursadlacin@gmail.com> Co-authored-by: kursad <kursad.lacin@brado.net> Co-authored-by: Dushyant Acharya <dushyantacharya873@gmail.com> Co-authored-by: Yuriy <yuriy.shuyskiy@gmail.com> Co-authored-by: Recep S <22618852+us@users.noreply.github.com> Co-authored-by: Moshe Malawach <moshe.malawach@protonmail.com> Co-authored-by: Moshe Malawach <moshemalawach@users.noreply.github.com> Co-authored-by: Rongkun Yan <2493404415@qq.com> Co-authored-by: Varshith <kvarshithgowda@gmail.com> Co-authored-by: Mateo Wang <277851410+mateo-berri@users.noreply.github.com> |
||
|
Sameer Kankute
|
bed6ce820c
|
test(batches): move orphan tests into tests/test_litellm for CI coverage (#30510)
Four batch-related tests lived under tests/litellm/ and were never picked up by GitHub Actions. Relocate them and fix gemini multimodal e2e to use the batchEmbedContents path expected for gemini/ provider. Co-authored-by: Cursor <cursoragent@cursor.com> |
||
|
Sameer Kankute
|
87cf67ec30
|
feat(gemini): forward web search tools in image generation (#30119)
* feat(gemini): forward web search tools in image generation Map tools and web_search_options to googleSearch on Gemini image generateContent requests for Google AI Studio and Vertex AI. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini): dedupe image search tools and return mapped params Skip web_search_options when tools already include search, dedupe search tool entries, and assign the return value from map_gemini_image_tools_params. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini): preserve toolConfig side-effects in image tool mapping * fix(gemini): forward toolConfig in image generation request body * fix(gemini): track web search grounding cost on image generation Forwarding Google Search grounding to Gemini and Vertex image generation previously incurred billable grounding charges that never reached LiteLLM spend tracking, because the image cost path returns through the Gemini/Vertex image calculators before built-in tool spend is added. Carry the grounding request count from the response onto the image usage object and bill it with the same per-request web search accounting used for chat completions. --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Sameer Kankute
|
cfcdf8714a
|
feat: litellm oss 110626 (#30202)
* Add gpt-realtime-whisper Realtime transcription support (OpenAI + Azure) (#29775) * Add gpt-realtime-whisper Realtime transcription support (OpenAI + Azure) Adds first-class support for the gpt-realtime-whisper streaming speech-to-text model, which uses the Realtime transcription session API rather than the file-based /audio/transcriptions path. Model registration: registers gpt-realtime-whisper and azure/gpt-realtime-whisper with audio-duration pricing (input_cost_per_second = 0.017/60, matching the published $0.017/minute input audio rate). REST endpoint: implements POST /v1/realtime/transcription_sessions (plus /realtime and /openai/v1 aliases) to mint an ephemeral transcription session for the WebRTC flow. Adds request/response types, OpenAI and Azure URL builders, a shared base handler (refactored from the client_secrets handler), the acreate_realtime_transcription_session SDK function, and route registration. The proxy encrypts the ephemeral key returned under client_secret.value and records the session type in the token so the follow-up /realtime/calls replays type=transcription rather than type=realtime. WebSocket: forwards intent=transcription through to the Azure handler (OpenAI already received it) with URL-encoding, so gpt-realtime-whisper opens a transcription session. Transcription-only sessions no longer trigger an erroneous response.create. Cost tracking: transcription sessions emit no response.done events; their usage arrives on conversation.item.input_audio_transcription.completed as {type: duration, seconds}. That usage is captured out-of-band (usage only, no transcript duplication) and billed by input_cost_per_second, with a token-billed fallback for token-priced transcription models. Adds tests for pricing math, URL builders, request/response types, the proxy route and SDK function, WebSocket intent forwarding, transcription-session streaming behavior, and the /realtime/calls session-type replay. * Address PR review: URL-encode all Azure WS query params; forward query_params through provider_config branch * Address PR review: session_type validation, model auth fix, cost perf, billing fallback, detail/docs cleanup * Improve test coverage: detection from backend, error paths, unknown usage type, resolved_model None * Backport realtime transcription websocket fixes * Enforce authorized realtime transcription model * Enforce realtime transcription model access * Enforce realtime resolved model scopes * Enforce WebRTC transcription model scope * Lazy evaluate debug log in pass-through endpoint (#30177) * Pass through debug lazy logging * fix(proxy): convert remaining eager pass-through debug logs to lazy formatting * fix(parallel_ai): migrate search integration from v1beta to v1 endpoint (#30157) * fix(parallel_ai): migrate search integration from v1beta to v1 endpoint The Parallel Search API moved from /v1beta/search (processor: base/pro, parallel-beta header) to /v1/search (mode: turbo/basic/advanced, no beta header). Request fields moved too: max_results, source_policy, and excerpt settings are now nested under advanced_settings, and source_policy uses include_domains/exclude_domains. The v1 response returns publish_date per result, which now maps to SearchResult.date instead of being hardcoded to None. The legacy processor param is mapped to the equivalent mode so existing callers keep working. * fix(parallel_ai): default mode to basic and simplify param handling The v1 API defaults to advanced mode when mode is omitted, while v1beta defaulted to the base processor. Without an explicit default, callers who pass no mode would be silently upgraded to a tier costing 2.25x more while litellm's cost map reports the basic-tier price. Sending mode=basic preserves the v1beta default and keeps cost tracking accurate. Also replaces the handled_params set with pop-as-consumed param handling so mapped params no longer need to be tracked in two places, and extends the tests to pin the default mode, processor=base mapping, mode-over-processor precedence, and top-level v1 param passthrough. * fix(parallel_ai): avoid double /v1 when api_base is already versioned A PARALLEL_AI_API_BASE like https://api.parallel.ai/v1 previously produced .../v1/v1/search. Strip a trailing /v1 before appending the search path and cover the api_base variants with a parametrized test. --------- Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> * feat(focus): add Mavvrik destination for FOCUS export (#29935) * fix: preserve responses streaming flag (#30189) * fix: preserve responses streaming flag * test: cover async responses streaming flag * fix(spend/daily-activity): stable offset pagination via id tiebreaker (#30164) (#30167) date alone is not a unique sort key for LiteLLM_DailyUserSpend or LiteLLM_DailyTeamSpend (many rows per date: api_key x model x model_group x provider x endpoint). Offset pagination over a non-unique sort landed on arbitrary boundaries, so a client paging through all results and summing per-page metrics (the Usage dashboard) got non-deterministic totals - sometimes inflated, sometimes deflated, different at different page_size values. Adding the row's UUID id (present on both tables) as a secondary sort gives every page a stable cursor. order=[{date desc}, {id asc}]. Fixes #30164 * fix(oci): inject a default maxTokens so omitted max_tokens doesn't truncate responses (#30018) * fix(oci): inject default maxTokens so omitted max_tokens doesn't truncate OCI GenAI applies a tiny server-side maxTokens default (~20 tokens) when the request omits it, so any call that doesn't send max_tokens comes back cut off mid-string with finishReason "length". MLflow judges never send max_tokens, so their JSON responses arrived as unterminated strings and json.loads failed in MLflow's gateway adapter. When no maxTokens/maxCompletionTokens target is set, inject DEFAULT_OCI_CHAT_MAX_TOKENS (env-overridable, defaults 4096), mirroring the Anthropic config's default-max-tokens behaviour. An explicit max_tokens still wins, and reasoning models still route to maxCompletionTokens. Used a fixed default rather than the catalog max_output_tokens because the catalog value is unreliable for some models (grok-4 reports max_output_tokens equal to its context window, not a real output cap, which would risk 400s). Adds TestOCIDefaultMaxTokens covering Cohere and generic injection, the explicit-override case, and the reasoning maxCompletionTokens branch. * test(oci): e2e regression that omitted max_tokens isn't truncated Real-proxy integration test asserting a chat completion that omits max_tokens completes with finish_reason "stop" instead of being cut off at OCI's ~20-token server default. Fails before the maxTokens-default injection (finish_reason "length", ~19 tokens), passes after. * test(oci): update cohere default-params test for injected maxTokens test_cohere_default_parameters asserted no maxTokens was injected, encoding the old behaviour where OCI's ~20-token server default truncated responses. Now that transform_request injects DEFAULT_OCI_CHAT_MAX_TOKENS, assert maxTokens equals that default while the other params (topK/topP/frequencyPenalty) stay pass-through with no hardcoded default. * fix(oci): make DEFAULT_OCI_CHAT_MAX_TOKENS a plain constant Drop the os.getenv override. The env knob was not requested and introducing a new env var forced a cross-repo dependency on litellm-docs (test_env_keys.py validates every referenced env var against the docs table there). A plain 4096 constant keeps the PR self-contained; callers who want a different limit pass max_tokens explicitly per request. * fix(oci): route all OpenAI commercial models to maxCompletionTokens OCI serves OpenAI models (gpt-4.1, gpt-5.1 through 5.5, o-series) that the litellm catalog doesn't track, so the supports_reasoning lookup returned False for them and the provider sent maxTokens, which the reasoning families reject with HTTP 400. With the injected default maxTokens this broke every request to those models, not just ones with an explicit max_tokens. Route the whole openai.* vendor prefix to maxCompletionTokens since OpenAI accepts max_completion_tokens on every chat model; the openai.gpt-oss-* open weights are served by OCI's own stack and keep maxTokens. Verified live against gpt-5.2, gpt-5, gpt-4o, gpt-4.1, gpt-oss-120b, llama-3.3, command-a and grok-3-mini * test(oci): hoist transformation imports and drop unused ones Makes the generic-chat test file ruff-clean: the per-test local imports of OCIChatConfig/OCIVendors shadowed the module-level import (F811) and left it unused (F401), and json plus three OCI type imports were never referenced * fix(oci): translate response_format json_schema to OCI's accepted shape (#29691) * fix(oci): translate response_format json_schema to OCI's accepted shape OCI GenAI rejected every json_schema response_format with HTTP 400 "Please pass in correct format of request", which broke structured-output callers such as MLflow LLM judges (they always send a json_schema). The provider forwarded OpenAI's raw json_schema body unchanged. For GENERIC models OCI's ResponseJsonSchema accepts only name/description/schema/isStrict, so OpenAI's `strict` key (and any other extra) 400s the request; the key must be renamed to isStrict and the body whitelisted. For Cohere models there is no JSON_SCHEMA type at all; the schema has to ride on JSON_OBJECT as {"type": "JSON_OBJECT", "schema": ...}. Cohere type values must also be the canonical uppercase TEXT/JSON_OBJECT. _normalize_response_format now branches by vendor and emits the exact shape each one accepts (verified live against OCI GenAI for Cohere, Meta, Gemini and Grok). Drops the unused, incorrect Cohere response-format pydantic models. Two existing tests asserted the broken behavior (lowercase type, raw jsonSchema on Cohere); they are rewritten to assert the corrected shape, and generic/Cohere json_schema regression tests are added. * fix(oci): raise early on json_schema response_format with no body A GENERIC model request with {"type": "json_schema"} and no json_schema object fell through to the JSON_OBJECT branch and emitted a bodyless {"type": "JSON_SCHEMA"}, which OCI rejects with an opaque HTTP 400. Raise a descriptive 400 at translation time instead. Cohere is unaffected since it always maps to JSON_OBJECT. * test(oci): gateway integration test for response_format json_schema Added to tests/integration/ (the real-network integration suite) reusing the existing OCI proxy harness, not tests/llm_translation/ which is mock-only. --------- Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix(oci): accept default n=1 on Cohere instead of hard-failing (#29705) * fix(oci): accept default n=1 on Cohere instead of hard-failing Cohere on OCI has no numGenerations field, so n was mapped to False and map_openai_params raised "param `n` is not supported on OCI" whenever a client sent n. But n=1 (and None) is the OpenAI default single-generation request, which every OCI model produces anyway, so standard clients that always send n=1 (such as the MLflow gateway) were rejected with a 500. Drop n=1/None silently for Cohere; only n>1 is genuinely unsupported and still raises (or drops under drop_params). Generic models are unaffected and keep numGenerations, including n>1. * docs(oci): explain why n is not advertised for Cohere despite tolerating n=1 * test(oci): gateway integration test for Cohere default n=1 Added to tests/integration/ (the real-network integration suite) reusing the existing OCI proxy harness, not tests/llm_translation/ which is mock-only. --------- Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix(oci): drop max_retries instead of hard-failing on OCI (#29727) max_retries is a litellm-level control param (litellm applies retries itself), not a generation param OCI accepts. The provider mapped it to False and raised "param `max_retries` is not supported on OCI" whenever it was present. The litellm proxy injects max_retries on every request, so any OCI call through the proxy 500'd unless drop_params was set. Drop max_retries silently in map_openai_params. Adds a unit test (Cohere and generic) and a gateway integration test that a plain request succeeds through a proxy without drop_params. Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix(spend-logs): rehydrate metadata JSONB text on ui_view_spend_logs (#29682) Fixes #29674. `/spend/logs/ui` raw-SQL path returns the JSONB metadata column as a string — prisma's query_raw skips the ORM-layer hydration. The UI reads metadata.status / metadata.error_information as object fields, so provider-failure rows look like successes. Fix: json.loads the metadata field right after query_raw, fall back to {} on malformed JSON. 3 existing error-code/error-message tests called json.loads on response.data[0]["metadata"] — they were leaning on the bug. Updated to read the dict directly. Plus 2 new regression tests (failure metadata roundtrip + invalid-json fallback). Reverting the fix makes both new tests fail with AssertionError: metadata should be dict, got <class 'str'>. * fix(proxy): release max_parallel_requests slot when a stream is cancelled mid-flight (#27955) (#30020) * fix(proxy): release max_parallel_requests slot when a stream is cancelled mid-flight (#27955) * fix: refund max_parallel_requests on disconnect from outer streaming generators The cancellation refund previously lived in async_post_call_streaming_iterator_hook, but that hook is nested inside the outer streaming generators and a nested async generator only receives GeneratorExit on garbage collection (non-deterministic). With only the v3 limiter enabled, /chat/completions also bypasses the hook entirely (needs_iterator_wrap() is false). Move the release into async_data_generator and async_streaming_data_generator, the generators Starlette closes on client disconnect, so the refund fires deterministically on every streaming route. Warn when no event loop is running, and document the window TTL refresh on the decrement * fix(mcp): propagate model into model_call_details for passthrough tool calls (#30122) * fix(mcp): propagate model into model_call_details for passthrough tool calls The @client decorator on call_mcp_tool creates the logging object via function_setup without a model kwarg, so model_call_details["model"] starts as None. execute_mcp_tool only set logging_obj.model as an instance attribute, which the spend-log writer never reads (it reads kwargs["model"] from model_call_details). MCP passthrough tools/call rows therefore persisted with model="" while list_tools rows showed "MCP: list_tools", degrading the Logs UI display and bucketing all MCP tool spend under an empty model in DailyUserSpend. Propagate the model into model_call_details alongside the existing attribute assignment so the StandardLoggingPayload and SpendLogs writer pick it up. Covers the /mcp passthrough, REST /mcp-rest/tools/call, and orchestrated paths (the latter already passed model into function_setup, so this is a no-op there). * test(mcp): trim regression test docstring * fix(mcp): surface upstream challenges for delegated OAuth (#30124) * fix(mcp): surface upstream challenges for delegated OAuth * docs(mcp): clarify delegated upstream auth comments * perf(benchmarks): add CPU timing metrics to streaming benchmark (#29980) * Add CPU timing metrics to streaming benchmark * Fix spacing around timing sample dataclass * fix(gemini): don't emit empty choices on metadata-only stream chunks (#29167) web_search + reasoning makes Gemini stream mid-chunks that carry only grounding/thought metadata — no content part, no finishReason. _process_candidates skips content-less candidates and the existing fallback only ran when finishReason was set, so choices stayed empty and the downstream streaming handler raised IndexError on choices[0]. Emit an empty-delta choice for content-less chunks regardless of finishReason. Fixes #28884 * fix(key): allow /key/update to clear budget_limits with [] or null (#30085) * Fix /key/update rejecting budget_limits clear requests with HTTP 400 Sending budget_limits: [] or null to /key/update returned HTTP 400, so once a key had budget windows the last one could never be removed. prepare_key_update_data only json.dumps'd budget_limits when the value was truthy, so [] and None passed through raw to the Prisma Json? column; jsonify_object only serializes dicts, and prisma-client-py has no DbNull sentinel for Json? writes, so Prisma rejected both shapes. Serialize the clear case explicitly as the JSON literal null, matching how memory_endpoints encodes metadata for the same column type. Truthy values keep the existing reset_at window initialization path. Fixes #30067. * Require admin access for budget_limits changes on /key/update Clearing budget_limits via [] or null is a budget mutation, but _validate_update_key_data only counted max_budget and spend as budget changes before deciding whether to skip _check_key_admin_access. A non-admin key owner or a team member with /key/update could therefore remove a key's per-window spend caps without admin authorization. Treat any explicit budget_limits value in the request (set, change, or clear) as a budget change so it gates through the same admin check as max_budget. model_fields_set is used because an explicit null is indistinguishable from an omitted field by value alone. * fix(proxy): persist guardrail info in spend logs for /v1/responses (#30092) Pre-call guardrail blocks on /v1/responses wrote guardrail_information as null in LiteLLM_SpendLogs because _handle_logging_proxy_only_error splits request_data by LoggedLiteLLMParams keys and litellm_metadata, where the Responses API stores request metadata including standard_logging_guardrail_information, was not among them. It fell into optional_params, so merge_litellm_metadata never saw it. Add litellm_metadata to LoggedLiteLLMParams so it routes into litellm_params the same way metadata does on the chat completions path Fixes #28971. * fix(proxy): handle non-standard SSE frames in Anthropic passthrough logging (#26000) Some third-party Anthropic-compatible providers emit non-standard SSE frames (OpenAI-style [DONE] sentinels, non-JSON keep-alive lines) in streaming responses. These caused json.JSONDecodeError in _build_complete_streaming_response, breaking the passthrough logging pipeline so the request was never logged or billed. Skip whole-line 'data: [DONE]' sentinels and catch JSONDecodeError per event. Matching the full line (not a substring) keeps a valid chunk whose text payload contains '[DONE]' from being dropped. Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Sameer Kankute <sameer@berri.ai> * feat(newrelic): Add New Relic extension (#26989) * initial New Relic integration. * Minor fixes for basic observability. * Implemented basic support for the success path. Generates New Relic custom events needed by the AI Monitorin interface. * Supportability metric is sent on first request. * Emit supportability metric every hour instead of once a day. * Add the start/end times to the messages before sending them so that the start time and end time reflect the correct time and both are not set to 'now'. * Make use of `turn_off_message_logging` configuration that is available by default from CustomLogger. * Enabling New Relic agent to be wired when docker container starts if an environment variable is set. * If we cannot find trace information, send the AI events without the trace ID attached. * Use a fake trace_id if we cannot find one. * Implementing a configuration so that users can use litellm configuration to disable sending LLM messages to New Relic. There is a second method to do this via New Relic env var. * Mised file. * Cleaning up logic to turn off recording content via either the LiteLLM configuration or an env var. * Removing debugging. Fixed logic / comments around how often to send supportability metric. * Initial version of public doc for New Relic. * Use a proper name for the doc file. * Updating newrelic.md document. * Updating LiteLLM documentation for New Relic extension. * Moving New Relic imports into the methods to support unit tests. * Adding unit tests for the New Relic extension. * Updating linting and the unit tests that are not running in the CI environment. * Address reviewer feedback on New Relic integration. - Fix _record_error_metric to use app.record_custom_metric() instead of module-level newrelic.agent.record_custom_metric() so the call works outside of an active transaction context - Remove unreachable except ImportError block in _get_trace_context - Update stale "23 hours" comment to "27 hours" (matches 97200s threshold) - Remove commented-out debug code from _process_success - Fix docs typo: NEW_RELIC_CUSTOM_INSIGHTS_EVENTS_MAX_SAMPLES_STOREDA -> NEW_RELIC_CUSTOM_INSIGHTS_EVENTS_MAX_SAMPLES_STORED - Update TestRecordErrorMetric to verify app.record_custom_metric call Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Reformating for the linter. * Addressing additional automated feedback. - Removed a legacy comment about the New Relic header - Reordered imports in one file - Switched another file to use the import at the top of the file instead of inline when used - Added unit tests for untested methods that were identified * Addressing new feedback. - Proper handling of time to floats. Created a util method and updated code to use it. - added the missing guard to ensure the app is enabled * Addressing feedback. - When an error occurs, still check if the periodic supportability metric should be emitted - Added a check to ensure the extension is ready in the error handler to match _process_success * Updating the NR event timestamps to more accurately reflect when the messages were generated. * Addressing feedback for potential better practice. * Addressing feedback on accessing default values. Added tests for most of these cases. * Adding a new catch exception block based on feedback. * Addressing feedback about a potential issue around a timestamp for the supportability metric. * Addressing minor feedback on length of generated, fallback traceId. * Addressing feedback. - A few more cases were found where the dictionary access might not return the correct value. - Handling cases where `traceparent` is not lower cased * Addressed feedback where the newrelic options might not apply correctly. * Addressing some feedback. * Addressing feedback. * Validating testing / formatting for our changes. * Updating linting, adding tests, defining data type for UI. * Configuration for the logging callback definition. * Adding a newrelic image for the UI to use. * Putting the New Relic callback in proper alphabetic order. * Copying the logo to a committed output directory so it shows up in a locally built container. * Adding missing definition of new env vars that were causing a build failure. * Addressing automated feedback from greptile. * Adding a few more unit tests to increase the code coverage just a bit more. * Additional unit tests to push coverage to almost 90%. * Adding a custom newrelic docker image build process. This removes the need to add the newrelic agent to the core litellm container or dependencies. * Clarifying message when the New Relic agent is not installed and someone is trying to use the newrelic extension. Either use the proper image when using docker, or install the agent manually when running from source. * Ensuring pip is available to install the New Relic agent. * Updating the definition and handling of traceId (no spanId). Clarifying behavior of env vars vs UI configuration for the newrelic extension. * Removing entries from the New Relic logger configuraiton UI as these values must be set as part of running the image. * Removing a stale doc file that has moved to the litellm-docs repo. Cleanup of Dockerfile to remove a LABEL that was incorrect. * Updating container image name to be the best guess for the new name. * Addressing feedback from greptile. - Added a comment around token_count=0 - Updated the boolean parser to allow a wider set of options which matches existing patterns in other parts of LiteLLM. * Removing option for a separate New Relic container image. The agreement is to handle this in the New Relic integration docs. * Updating error message when New Relic agent is not available. * Wiring in the test message from the LiteLLM callback UX. * Missed saving one of the file conflicts. * Fixed a lint error I introduced. Somehow, I dropped another string and now added it back. * Adding newrelic to the schema definition. * Added an admin check on the call before sending test message as mentioned by the AI code review. * Updating to use should_redact_message_logging(kwargs) as part of the logic to determine if message content should be sent to New Relic or not. This still uses the `record_content` property as well, but both have to be true in order for content to be included. --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> * Add Azure AI Foundry DeepSeek V3.1 and V4 Pro/Flash global pricing to cost map (#30134) Co-authored-by: Cursor <cursoragent@cursor.com> * fix(logging): translate Responses bridge result to ModelResponse for spend logs (#28985) PR #29394 fixed the AnthropicResponse.model_validate crash for the streaming anthropic_messages -> OpenAI Responses bridge by unwrapping terminal events and returning the inner ResponsesAPIResponse. The spend_logs row lands and usage/cost are correct, but the row's response field stores the Responses API shape (output[...].content[...].text). The proxy UI Logs tab reads response.choices[0].message via parseMessages in prettyMessagesUtils.ts with no fallback for the Responses shape, so the OutputCard renders "No response data available" for every cross-routed call. The same shape mismatch affects every downstream consumer of spend_logs that assumes the canonical chat-completion shape This change keeps the unwrap from #29394 but routes the resulting ResponsesAPIResponse (and the bare-response non-streaming path) through LiteLLMResponsesTransformationHandler.transform_response, which is the same conversion already used by the chat-completion Responses bridge. Spend_logs now stores a ModelResponse with choices[0].message.content, so the UI and other consumers see the assistant text. On a translation failure (eg. empty output on an incomplete response) the handler falls back to a minimal ModelResponse carrying model and usage so the row still lands rather than being dropped as a Non-Blocking error Also corrects a stale comment in the Responses adapter that implied the call type was reclassified to acompletion; the code preserves anthropic_messages and the success handler translates back to ModelResponse for the row Fixes #28595 * fix(anthropic-adapter): re-emit first delta on streaming content-block transitions (#30024) * fix(anthropic-adapter): re-emit first delta on streaming content-block transitions The `/v1/messages` -> `/v1/chat/completions` streaming adapter (`AnthropicStreamWrapper`) silently dropped the first non-empty delta of every content block that started via a *transition* (e.g. text -> tool_use -> text, text -> thinking). When an upstream chunk both triggers a new content block (its type differs from the active block) and carries that block's first delta, the wrapper emitted `content_block_stop` -> `content_block_start` and then only re-queued the trigger chunk when it was an `input_json_delta` (bundled tool args). The synthesized `content_block_start` always carries an empty body, so the first `text_delta` / `thinking_delta` was lost — the client output started from the second token (e.g. "Hi, how can I help you?" rendered as ", how can I help you?", or text resuming after a tool call lost its first sentence). This is especially visible with Claude Code-style clients that consume Anthropic Messages streaming events strictly. Fix: re-queue the trigger chunk's translated delta whenever it carries non-empty content (text/thinking/signature/tool args), via a shared `_trigger_delta_has_content` helper used by both the sync and async paths. Empty trigger deltas are still suppressed so no spurious empty `content_block_delta` is introduced. Fixes #30014 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * test(anthropic-adapter): cover all _trigger_delta_has_content branches Add a direct parametrized unit test for the re-emit predicate so every delta type (text/input_json/thinking/signature), the empty-payload guards, and the malformed/non-delta cases are exercised independently of upstream chunk translation. Raises patch coverage for the new helper. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> --------- Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> * feat: add opt-in healthy_only filter to GET /v1/models (#30130) * feat: add opt-in healthy_only filter to GET /v1/models Adds an opt-in `healthy_only=true` query parameter to GET /v1/models and GET /models that hides models whose backing deployments are all marked unhealthy by background health checks. - Add Router.async_get_fully_unhealthy_model_names(), mirroring the semantics of get_fully_blocked_model_names(): a model is hidden only when every backing deployment is unhealthy and the health state is not stale (fail open otherwise). - Reuses the existing DeploymentHealthCache populated by _run_background_health_check(), so no new health state is introduced. - No-op when allowed_fails_policy is set, mirroring _async_filter_health_check_unhealthy_deployments semantics. - team_public_model_name aliases are aggregated alongside model_name. - Hiding is presentation-only; default behavior is unchanged. Fixes #30128 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * docs: address Greptile review notes - Note team-alias asymmetry vs get_fully_blocked_model_names - Debug-log when healthy_only is set but no health state is available Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> --------- Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Claude Fable 5 <noreply@anthropic.com> * Dedupe team soft budget alerts by team_id instead of token (#30097) _team_soft_budget_check sends type="soft_budget" alerts with event_group=TEAM, but SoftBudgetAlert.get_id always returned the request token. The alert cache key was therefore scoped per virtual key, so every active key in a team over its soft budget fired its own alert within budget_alert_ttl. Branch on event_group so team-level alerts dedupe by team_id, matching TeamBudgetAlert, while key and project level alerts keep per-token dedupe. Fixes #27398. * feat(bedrock guardrails): support contextual grounding qualifiers (request-side) (#30057) * test: add failing tests for Bedrock contextual grounding (request-side) Drive the request-side of Bedrock contextual grounding: callers tag message content blocks as grounding_source/query, the post_call hook assembles an ApplyGuardrail(OUTPUT) call carrying source + query + response(guard_content), and the bedrock converse transform must render the tags as prompt text instead of silently dropping them. Non-grounding payloads must stay byte-identical. * feat(bedrock guardrails): support contextual grounding qualifiers Bedrock contextual grounding scores a model response against a reference source and the user query, expressed via a per-content-block `qualifiers` array on ApplyGuardrail. The guardrail hook previously sent plain text only, so grounding could not be driven through it even though the response-side contextualGroundingPolicy parsing already existed. Callers now tag message content blocks `{"type":"grounding_source"}` / `{"type":"query"}` (mirroring the existing `guarded_text` marker). On the generate path the bedrock converse transform renders them as plain text; at post_call the hook harvests them from the request and assembles one ApplyGuardrail(OUTPUT) call carrying grounding_source + query + the response (as guard_content). Requests without these tags produce a byte-identical payload, so existing behaviour is unchanged. * Feat(guardrail): Adding support for custom Ovalix guardrail (#21887) * Feat(guardrail): Adding support for custom Ovalix guardrail * Internal CR comments fixes * greptileai comments fixes * fix conflict * fixes * fix sha256 * clarify Ovalix actor-id hash is for normalization, not PII protection * fix(github_copilot): normalize per-event item_id in /responses streaming (#30072) GitHub Copilot's native /v1/responses stream assigns a different item_id to every event of a single output item (output_item.added, the part.added / delta / done events, and output_item.done). Spec-strict clients like the Vercel AI SDK key streaming parts by item_id and abort with "reasoning part <id> not found" / "text part <id> not found" when a delta references an unregistered id. Override transform_streaming_response in GithubCopilotResponsesAPIConfig to anchor every event of an output item to the id from its output_item.added. Copilot accepts that id paired with the final encrypted_content on the next turn, so multi-turn replay is unaffected. Fixes #30071 * feat: add /model/block and /model/unblock endpoints (#30125) * feat: add /model/block and /model/unblock endpoints Add dedicated proxy-admin POST /model/block and /model/unblock endpoints over the existing blocked flag on LiteLLM_ProxyModelTable, mirroring the /key/block and /key/unblock pattern. Calling a model whose deployments are all blocked now returns a clear 403 "Model is blocked" instead of a generic no-deployment error, including direct-dispatch route types (e.g. eval) via a pre-route guard. Includes audit-log entries for block/unblock and unit tests. Closes #29742 Signed-off-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> * chore: regenerate dashboard API types for model block/unblock endpoints Regenerate ui/litellm-dashboard/src/lib/http/schema.d.ts from the proxy OpenAPI spec (npm run gen:api) so it includes the new endpoints. Signed-off-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> * fix: widen router block-helper param type and add direct unit tests Type the _are_all_deployments_blocked deployments parameter to match its callers (DeploymentTypedDict) so mypy passes, and add tests/test_litellm/test_router_block_helpers.py with direct unit tests for the three block helper methods so router_code_coverage recognizes them. Signed-off-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> * fix: restore type-ignore on messages arg after black reflow Signed-off-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> * refactor: raise model-block 403 in proxy layer, not SDK Router Keep the SDK Router's documented behavior for blocked deployments (filtered -> "no healthy deployment") and move the 403 PermissionDeniedError into the proxy layer (route_llm_request), where model blocking is an admin concept. This avoids a backwards-incompatible 403 for SDK users who set blocked=True on their own deployments, per maintainer review. Signed-off-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> --------- Signed-off-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> Signed-off-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> Co-authored-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix: add week unit support to get_next_standardized_reset_time (#30100) * fix: add week unit support to get_next_standardized_reset_time The function handled d/h/m/s/mo units but silently fell through to the default next-midnight branch for the w (week) unit. This was inconsistent: _extract_from_regex already accepted w in its character class, and duration_in_seconds already returned value * 604800 for it. Add the missing elif unit == 'w' branch that delegates to _handle_day_reset with value * 7, which reuses the existing Monday- alignment logic for 1w and the generic N-day-from-midnight path for larger multiples. Add test_week_based_resets covering 1w from a Wednesday (expects next Monday) and 2w from a Monday (expects 14 days forward at midnight). Signed-off-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> * test: exercise relative week semantics with non-Monday base dates + add docstring Signed-off-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> --------- Signed-off-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> Co-authored-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> * fix: black formatting and remove undocumented MAVVRIK_FOCUS_FREQUENCY env var * fix: black formatting with correct version and sync schema.d.ts for healthy_only param * fix: resolve mypy errors and add transcription_sessions to JSON schema endpoint enum * fix: restore MAVVRIK_FOCUS_FREQUENCY guard and exclude it from docs key scan * fix: address Greptile P2 comments - move constant, use UTC datetime, skip redundant team lookup * revert: restore original team lookup logic in can_key_call_resolved_model --------- Signed-off-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> Signed-off-by: FugoP <264910004+AgentGymLeader@users.noreply.github.com> Co-authored-by: Emerson Gomes <emerson.gomes@thalesgroup.com> Co-authored-by: nina-hu <nina.huuu@gmail.com> Co-authored-by: Sahith Jagarlamudi <104647530+s-jag@users.noreply.github.com> Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Praveen Ghuge <95286176+pghuge-cloudwiz@users.noreply.github.com> Co-authored-by: alex107ivanov <30668368+alex107ivanov@users.noreply.github.com> Co-authored-by: hcl <chenglunhu@gmail.com> Co-authored-by: Fede Kamelhar <federico.kamelhar@oracle.com> Co-authored-by: Armaan Sandhu <74664101+Ar-maan05@users.noreply.github.com> Co-authored-by: Teo Xian Zhong Augustine <35527068+auggie246@users.noreply.github.com> Co-authored-by: King Star <mcxin.y@gmail.com> Co-authored-by: Saksham Maggo <122939011+SakshamMaggo@users.noreply.github.com> Co-authored-by: Filippo Menghi <113345637+Cyberfilo@users.noreply.github.com> Co-authored-by: Kelvin <leikaiwei@outlook.com> Co-authored-by: Josh Bonczkowski <josh.bonczkowski@gmail.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> Co-authored-by: M. Dennis Turp <mdturp@pm.me> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: Piotr Minkina <piotrminkina@users.noreply.github.com> Co-authored-by: Martín Alcalá Rubí <martin@tryolabs.com> Co-authored-by: T. Kobayashi <13004314+nix-tkobayashi@users.noreply.github.com> Co-authored-by: João Costa <13508071+jpv-costa@users.noreply.github.com> Co-authored-by: Shalom <shalom@ovalix.io> Co-authored-by: codgician <15964984+codgician@users.noreply.github.com> Co-authored-by: FugoP <kim@pomsora.com> Co-authored-by: AgentGymLeader <264910004+AgentGymLeader@users.noreply.github.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Sameer Kankute
|
3b40ac987f
|
Litellm oss 090626 (#30021)
* fix(mcp): report scoped server name during initialize (#29865) * fix mcp scoped server name * Update litellm/proxy/_experimental/mcp_server/mcp_context.py Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * test(mcp): cover scoped server name in the SSE initialize handler --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(ui): show all session logs in the drawer, not just the first 50 (#29795) * fix(ui): show newest session logs first * test(ui): keep session log pagination coverage * fix(ui): show all session logs in the drawer, not just the first page The session detail drawer fetched session logs via sessionSpendLogsCall without page/page_size, so it only ever received the backend default of one page (50 rows). Sessions with more than 50 calls had the rest unreachable in the UI (#29153). sessionSpendLogsCall now takes page/page_size, and the drawer fetches the first page, reads total_pages, then fetches the remaining pages and accumulates them before the existing client-side sort. This keeps the single continuous list (and the selected-log lookup and keyboard navigation, which all assume the full session) correct. Fetching is bounded by a page cap, and the sidebar shows a "showing most recent N" note if a session exceeds it. The rows are lightweight metadata (the endpoint excludes messages/response), so the full set is small; request/response bodies are still loaded per log on demand. * fix(ui): default session drawer to most recent log, newest first Open a session with its most recent log selected, and order the sidebar newest-first to match the all-sessions logs overview. MCP calls stay grouped last. The latest log by time is computed explicitly, since the MCP grouping means it is not always the first row. * Apply fetching pages in batches suggestion from @greptile-apps[bot] Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(ui): derive session total from accumulated rows when backend omits it Compute the session total after all pages are fetched, falling back to the accumulated row count rather than the first page's. Guards the truncation note against a backend response that omits total but spans multiple pages. --------- Co-authored-by: Yufeng He <40085740+he-yufeng@users.noreply.github.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(proxy): handle Mistral multipart passthrough (#29927) * fix(proxy): handle Mistral multipart passthrough * chore: satisfy passthrough ci formatting * test(proxy): cover Mistral passthrough in CI shard * fix(vertex_ai): use REP host for context caching on eu/us multi-region endpoints (#29573) Context caching built the cachedContents URL as https://{location}-aiplatform.googleapis.com, which is an invalid host for the eu/us multi-region endpoints and returns 404. The inference path already resolves these to the REP host (https://aiplatform.{geo}.rep.googleapis.com) via get_vertex_base_url(); reuse that helper in _get_token_and_url_context_caching so caching uses the same host as inference. Adds tests covering the eu/us multi-region cachedContents URLs (v1 and v1beta1). Fixes #29571 * Support per-model encrypted content affinity config (#29760) Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> * fix: propagate upstream status code in proxy API exception handler (#29402) * fix: propagate upstream status code in proxy API exception handler When Google GenAI / Vertex returns a 404 for deprecated or missing models via streamGenerateContent, the exception was falling through to a generic handler that defaulted to 500. Now provider exceptions carrying a valid HTTP status_code correctly propagate it through to the ProxyException. * fix: apply black formatting to common_request_processing.py * fix: tighten status code range to 400-599 and deduplicate ProxyException raise * fix(tests): use valid vertex_location in context caching tests Replace "test_location" (contains underscore) with "us-central1" so tests pass the regex validation added in get_vertex_base_url(). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat(sdk): add xAI OAuth provider (#29866) * Add xAI OAuth provider * Update oauth.py Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * Fix xAI OAuth CI failures * Add xAI OAuth coverage tests * Move xAI OAuth coverage tests to core utils * Address xAI OAuth review comments * Prevent xAI OAuth api_base token exfiltration * Treat blank xAI OAuth api keys as absent * Wrap invalid xAI OAuth JSON responses * Use xAI OAuth behind explicit flag --------- Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(proxy) #27734 allow clearing budget_duration and team_member fields by sending null on /key/update and /team/update (#27751) * fix(proxy): allow clearing budget_duration and team_member fields by sending null on /key/update and /team/update Fixes #27734 Sending null for budget_duration, team_member_budget, team_member_budget_duration, team_member_rpm_limit, or team_member_tpm_limit via /key/update or /team/update returned 200 OK but silently ignored the null value. The fields remained unchanged in the database. Root causes: - /key/update: prepare_key_update_data() popped budget_duration from the update dict but never re-added it (or budget_reset_at) when the value was None. - /team/update: _set_budget_reset_at() only acted when budget_duration was non-None, leaving a stale budget_reset_at in the DB. - /team/update: team_member_* null values bypassed the budget table update entirely because should_create_budget() requires at least one non-None field. * test(proxy): cover no-budget-row path in clear_team_member_budget_fields * fix(presidio): unmask PII tokens in Anthropic native SSE streaming bytes (#30028) * fix(presidio): unmask PII tokens in Anthropic native SSE streaming bytes When output_parse_pii=true on the Anthropic native path (anthropic/claude-*), response chunks arrive as raw bytes in SSE format. _stream_pii_unmasking was yielding those bytes unchanged, so <PERSON_1> tokens were never replaced with the original values before reaching the caller. Add _unmask_sse_bytes_chunk to parse each data: line, find content_block_delta / text_delta events, and apply _unmask_pii_text before re-encoding. Wire it into _stream_pii_unmasking so bytes chunks are unmasked when pii_tokens exist. * fix(presidio): handle CRLF line endings and non-ASCII PII in SSE unmask Strip trailing \r before the [DONE] guard so CRLF-terminated SSE chunks don't bypass it and silently swallow a JSONDecodeError. Add ensure_ascii=False to json.dumps so non-ASCII replacement values like accented names are preserved as UTF-8 on the wire rather than being \uXXXX-escaped. Add regression tests for both cases. * feat(bedrock_mantle): path-aware Responses routing (/v1/responses vs /openai/v1/responses) (#29925) * feat(bedrock_mantle): path-aware Responses routing (/v1/responses vs /openai/v1/responses) Bedrock Mantle serves the Responses API on two upstream paths: - gpt frontier models (gpt-5.5 / gpt-5.4) on /openai/v1/responses - every other Responses-capable model (e.g. gpt-oss) on the standard /v1/responses BedrockMantleResponsesAPIConfig gains a `use_openai_path` flag; the provider gate in utils.py picks the path per model: openai.gpt-* (non gpt-oss) -> /openai/v1/responses; any model declared mode=responses (price-map entry or user model_info) -> /v1/responses; everything else returns None and keeps the existing chat-completions emulation. Adds gpt-5.5 / gpt-5.4 price-map entries, registry wiring, and the routing-matrix tests. * feat(bedrock_mantle): data-driven frontier routing via use_openai_responses_path Addresses the Greptile review point that frontier detection should be a price-map field rather than a hardcoded name match. The gate now routes a model to /openai/v1/responses when its price-map entry declares use_openai_responses_path, so a frontier model whose name does not follow the openai.gpt- convention can be onboarded by JSON alone. The name-convention check is kept as a fallback that needs no price-map entry, which preserves zero-change routing for a future gpt-6 before its entry loads. gpt-5.5 / gpt-5.4 get the flag in both price maps. Adds tests for the data-driven flag path and for the flag presence on the gpt-5.x entries; both branches are mutation-tested. * test(model_prices): allow use_openai_responses_path in price-map schema The model_prices_and_context_window.json schema validator (test_aaamodel_prices_and_context_window_json_is_valid) enforces additionalProperties: false, so the new use_openai_responses_path flag on the gpt-5.5 / gpt-5.4 entries failed validation. Add it to the schema as a boolean, alongside the other supports_* / capability flags. * Add Tensormesh serverless models to the model cost map (#30037) * Add Tensormesh serverless models to the model cost map * Flag reasoning support on the Tensormesh models that expose thinking mode * fix(proxy): invalidate stale key spend counter after budget reset or manual spend update (#30001) * fix(proxy): reconcile stale key spend counter after budget reset * fix(proxy): invalidate stale key spend counter after budget reset or manual spend update * fix(proxy): remove read-time stale counter reconciliation to prevent budget bypass * revert: undo unrelated formatting changes in enterprise directory * test(proxy): add unit test for key spend update invalidating counter * test(proxy): fix mocked update_data and hash token expectations in unit test * fix(proxy): use Responses-API transformer in pass-through cost tracking (#29728) The `elif is_responses:` branch of `openai_passthrough_handler` was calling the chat-completions `transform_response` on a Responses API payload. The chat-completions transformer expects `choices: [...]` in the raw response; the Responses API uses `output: [...]` and `usage.input_tokens` / `usage.output_tokens` (not `prompt_tokens` / `completion_tokens`). The result was a KeyError 'choices' deep inside `convert_to_model_response_object`, swallowed by the surrounding `except Exception` in the handler, and the SpendLogs row was written by the fallback path with zeroed-out tokens, spend, and model. This bug silently undercounts cost for every successful pass-through call to either OpenAI's `/v1/responses` or Azure's `/openai/v1/responses` (deployments configured for the Responses API). Reproduced 2026-06-04 against a real Azure OpenAI Responses API deployment proxied through LiteLLM v1.88.0. Fix: use the dedicated `OpenAIResponsesAPIConfig.transform_response_api_response` for the Responses branch. This transformer already exists in LiteLLM (`litellm/llms/openai/responses/transformation.py`) and knows the Responses-API on-the-wire shape. `litellm.completion_cost` already handles `ResponsesAPIResponse` natively with `call_type="responses"`, so no downstream changes are needed. Tests: test_responses_api_uses_responses_transformer_not_chat_completions NEW. Real regression test — exercises the openai_passthrough_handler with a real-shaped Responses payload (no `choices`, has `output` and Responses-API `usage` keys) and NO mocked `get_provider_config`. Pre-fix: raises KeyError 'choices' inside the chat-completions transformer (the bug). Post-fix: returns a ResponsesAPIResponse, completion_cost is called with call_type="responses" and a ResponsesAPIResponse instance (asserted). Verified to fail on un-fixed handler + pass on fixed handler before commit. test_responses_api_cost_tracking UPDATED. Old test mocked `get_provider_config` (no longer called in the responses branch post-fix). Now mocks the Responses transformer directly (`OpenAIResponsesAPIConfig.transform_response_api_response`) to test the downstream cost-calc contract. Out of scope for this PR (separate followup): - Recognizing *.cognitiveservices.azure.com (the newer Azure OpenAI hostname) in the is_openai_*_route checks. Separate PR. Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> * fix(skills): execute DB skills by matching the litellm_skill_ tool name prefix (#30116) Skill IDs are generated as litellm_skill_<uuid> and the model-facing tool name is the sanitized skill ID, but the post-call execution gates in SkillsInjectionHook only ran tools whose name starts with "skill_", so DB skills were silently returned to the client as raw tool calls. Fixes #28122. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(anthropic): synthesize content_block_start when Responses stream omits output_item.added (#30115) * fix(team): reserve team budget raises for proxy admins on /team/update (#30030) The caller's PERSONAL max_budget was the wrong yardstick for /team/update: a team's spend ceiling has nothing to do with the admin's own key budget. That comparison was an unintended side effect of reusing _check_user_team_limits() (which exists for the /team/new path) and broke the UI, which re-sends the unchanged budget on every save. New behavior on /team/update for standalone teams: - A team admin (already authorized via _verify_team_access) may freely KEEP or LOWER the team budget, and change models/tpm/rpm, without being gated by their personal limits. - GROWING a team's spend ceiling is a budget-authority action reserved for proxy admins -> 403 for team admins. "Growing" covers both raising max_budget above the team's current finite value and removing the cap entirely (max_budget=null, detected via model_fields_set so an explicit null is distinguished from an omitted field). For a team that currently has no cap, setting a finite value is a restriction and is allowed. - Org-scoped teams remain governed by _check_org_team_limits() (capped by the org budget). Also reverts the #29525 existing_team_max_budget workaround in _check_user_team_limits() back to the create-only form; /team/new still enforces the creator's personal caps. docs(access_control): resolve the contradiction in the team-admin section — team admins can keep/lower the budget and manage rate limits/models, but cannot raise the team budget (proxy-admin only). tests: unit + behavior coverage for raise-blocked, cap-removal-blocked (team admin), raise/removal allowed (proxy admin), uncapped-team restriction allowed, keep/lower/resend allowed, and unchanged create-path guards. Co-authored-by: Cursor <cursoragent@cursor.com> * test(ui): data-driven App Router migration E2E smoke (default + server-root-path) (#29974) * test(ui): add a data-driven App Router migration E2E smoke Add a growing Playwright smoke for migrated pages: for each segment it deep-links to the path route, asserts the URL and that the dashboard shell rendered, then clicks off to a legacy page and asserts navigation still works. Driven by e2e_tests/fixtures/migratedPages.ts, so adding a page is one line. Runs in two situations against the same proxy: the default mount (npm run e2e:migration) and a non-root SERVER_ROOT_PATH mount (npm run e2e:migration:root). globalSetup now logs in at `${SERVER_ROOT_PATH}/ui/login` so the admin storage state is valid under a prefix. Seeded with api-reference; append the rest as their migrations merge. * test(ui): support headed slow-motion + watch pauses in the migration smoke Honor SLOWMO in the server-root-path config (the default config already did), and add an env-gated E2E_WATCH_MS pause so a headed run lingers on each state. Both are no-ops by default, so CI behavior is unchanged. * test(ui): make the migration smoke a sidebar-click user journey Rework the smoke from deep-linking to a real navigation journey: start at the landing page, click the migrated page in the sidebar (expanding submenus for nested items), assert the path route rendered, reload it (the check a wrong server_root_path breaks), bounce to a legacy page and back, and — once two pages are migrated — navigate directly between two migrated pages. Verifies via URL + shell render, driven by the same fixture list. * test(ui): address review on the migration smoke Escape ROOT and segment before interpolating them into RegExp URL matchers so a future segment containing regex metacharacters can't silently widen the match. Make the server-root-path config fail fast when SERVER_ROOT_PATH is unset instead of silently re-running the default mount and passing without exercising the prefix. * test(ui): drop unused watch helper and fix stale smoke README * test(ui): run the migration smoke under a server root path in CI * test(ui): harden + instrument the server-root-path proxy reboot in CI * test(ui): run the server-root-path migration smoke as its own CI job Replace the in-place proxy reboot in e2e_ui_testing with a dedicated e2e_ui_testing_server_root_path job that boots the proxy once with SERVER_ROOT_PATH=/litellm, matching how every other proxy variant in the config gets its own job rather than killing and relaunching the live proxy. The reboot was failing deterministically: after pkill -9 and relaunch the prefixed proxy never came back up on :4000 (connection refused), so the smoke never ran. The readiness step that was supposed to surface the cause could never reach its boot-log tail because CircleCI runs steps under bash -eo pipefail and the preceding `curl -sv ... | tail` aborted the step with curl's exit 7. Booting the proxy as the job's own background step lets any boot crash land in that step's log instead of being swallowed. The default e2e_ui_testing job is unchanged aside from dropping the reboot, prefixed-readiness, and prefixed-smoke steps; the migration smoke still runs at the root mount there via the default Playwright config. * fix(proxy): extend response headers hook to streaming, TTS, image gen, and pass-through (#24232) * fix(proxy): extend response headers hook to streaming, TTS, image gen, and pass-through * test: mock post_call_response_headers_hook in audio speech route tests * chore(ui): remove dead App Router route stubs under (dashboard) (#30045) models-and-endpoints, organizations, and virtual-keys each had a page.tsx route under (dashboard)/ that is not in MIGRATED_PAGES, so the sidebar and deep links never resolve to it and the route is unreachable. Each was a thin wrapper that handed the shared view empty or no-op props (empty modelData with a no-op setModelData, hardcoded empty organizations, no-op setUserRole/setUserEmail), so reaching one would render a degraded page in any case. The real wrapper belongs in the PR that flips each page into MIGRATED_PAGES, written with eyes on it and a test This continues the dead-scaffolding cleanup from #28891. The shared components these wrappers rendered (ModelsAndEndpointsView, OrganizationFilters) stay, since the legacy ?page= switch in app/page.tsx and src/components still import them * fix(ui/mcp): reset OAuth state on create-server modal close so a prior server's token no longer leaks into the next add-server session (#30000) * fix(ui/mcp): reset OAuth hook state on modal close so a prior server's token no longer leaks into the next add-server session * fix(ui/mcp): clear in-flight OAuth guard on reset and reset form/tools on modal close so nothing leaks on a parent-driven dismiss * fix(mcp): allow team access-group grants in OAuth authorize/token access check (#30041) * fix(mcp): honor team access-group grants in OAuth authorize/token access check * test(mcp): mock build_effective_auth_contexts in non-admin authorize tests for isolation * docs(security): require a reproduction video for vulnerability reports (#30048) (#30063) With AI models capable of automated vulnerability discovery now publicly available, we expect a large increase in report volume, much of it unverified. Requiring a video of the exploit running against a live instance raises the bar for submissions and keeps triage focused on reproducible issues. Reports without a video will be closed and reopened if one is added later. Co-authored-by: stuxf <70670632+stuxf@users.noreply.github.com> * feat(ui): add admin flag to disable in-product UI nudges for everyone (#29796) * feat(ui): add admin flag to disable in-product UI nudges for everyone Admins can now suppress the survey and Claude Code feedback popups for all users via a single disable_ui_nudges UI setting, instead of relying on each user dismissing them individually. * fix(ui): suppress nudges while ui settings are loading Gate nudgesDisabled on the ui-settings loading state so an admin with disable_ui_nudges on doesn't see the survey prompt flash, and the getInProductNudgesCall fetch doesn't fire, on a cold page load before the flag resolves. Falls back to showing nudges if the fetch errors. * test(ui): wrap CreateKeyPage test in QueryClientProvider page.tsx now calls useUISettings (react-query), which needs a QueryClient that layout.tsx supplies in production but the test did not. Add the provider and mock getUiSettings so the query resolves. * chore(ui): remove dead dashboard files and unused dependencies (#30047) * chore(ui): remove dead dashboard files and unused dependencies knip flagged seven orphaned source/config files with no importers and five declared dependencies that nothing in the tree uses. Removing them shrinks the dashboard bundle's source surface and keeps the manifest honest; vite stays installed transitively via vitest, so test tooling is unaffected. * fix(ci): restore serverRootPath.config.ts referenced by SERVER_ROOT_PATH workflow The dead-code sweep removed e2e_tests/serverRootPath.config.ts, but its spec (tests/login/serverRootPathRedirect.spec.ts) and the test_server_root_path.yml workflow step still depend on it, so the redirect e2e job failed to load a config that no longer existed. * fix(proxy): authorize batch files using upload target_model_names (LIT-3593) (#30009) * fix(proxy): authorize batch files using upload target_model_names (LIT-3593) After replace_model_in_jsonl, body.model is a stripped provider id. Reverse-mapping it via resolve_model_name_from_model_id is first-match on model_list and caused false 403s when multiple deployments share the same stripped name. Use target_model_names from the unified file id instead. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(proxy): restore resolve_model_name_from_model_id for JSONL fallback path (LIT-3593) Restores the reverse-lookup for the JSONL body.model fallback path so that legacy/pre-target_model_names managed files still map stripped provider IDs back to proxy aliases before auth. Also cleans up redundant `or None`. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * Revert "fix(proxy): restore resolve_model_name_from_model_id for JSONL fallback path (LIT-3593)" This reverts commit |
||
|
Sameer Kankute
|
dfd6cbc514
|
fix(vertex): propagate Vertex AI metadata in streaming success callbacks (#29899)
* fix(vertex): propagate Vertex AI metadata in streaming success callbacks Streaming calls assembled via stream_chunk_builder were missing vertex_ai_grounding_metadata and vertex_ai_url_context_metadata in standard_logging_object.response. Merge metadata from chunks into the assembled response and mirror non-streaming hidden_params on Gemini chunks. Co-authored-by: Cursor <cursoragent@cursor.com> * refactor(vertex): move streaming metadata merge into provider config hook Address review feedback by delegating assembled-stream metadata propagation to VertexGeminiConfig via BaseConfig.apply_assembled_streaming_response_metadata, and only write chunk hidden_params when metadata is non-empty. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(redaction): scrub Vertex provider metadata when message logging is off Clear vertex_ai_grounding_metadata and related fields from standard logging responses and assembled streaming ModelResponse objects so turn_off_message_logging cannot leak prompt-derived web search queries. Co-authored-by: Cursor <cursoragent@cursor.com> * Use assembled model for streaming metadata hook * Fix Vertex metadata redaction bypass in logging callbacks. Scrub Vertex provider fields from litellm_params.metadata.hidden_params during perform_redaction so streaming success_handler merges do not leak prompt-derived metadata when message logging is disabled. Co-authored-by: Cursor <cursoragent@cursor.com> * Fix Vertex streaming metadata from hidden params * fix(vertex): mirror vertex_ai_safety_results on assembled streaming responses The non-streaming transform_response stores safety data under vertex_ai_safety_results, but the streaming path only wrote vertex_ai_safety_ratings. Assembled streaming responses therefore never carried vertex_ai_safety_results, so any consumer reading that field saw a silent difference between streaming and non-streaming calls. Set vertex_ai_safety_results alongside vertex_ai_safety_ratings in the shared stream metadata setter and add it to the assembled metadata field list so it propagates through stream_chunk_builder. * fix(streaming): log provider streaming metadata hook failures instead of swallowing them * refactor(vertex): share single Vertex metadata field tuple across redaction and streaming * refactor(vertex): move Vertex metadata redaction helpers into llms/vertex_ai --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Sameer Kankute
|
cb041966bf
|
Litellm oss staging 040626 (#29671)
* fix(azure): apply api_version fallback chain to image edit URL
`AzureImageEditConfig.get_complete_url` only read `api_version` from
`litellm_params`. When callers configured it via `litellm.api_version`
or `AZURE_API_VERSION`, the constructed URL had no `?api-version=` and
Azure responded `404 Resource not found`.
Apply the same fallback chain the Azure chat path already uses in
`common_utils.py`:
litellm_params > litellm.api_version > AZURE_API_VERSION env >
litellm.AZURE_DEFAULT_API_VERSION
Adds 5 unit tests pinning each layer of the chain plus a regression
guard for `api_base` that already carries `?api-version=`.
* feat(mcp): core sampling and elicitation flow with security hardening
- Add sampling_handler.py: full MCP sampling/createMessage flow with
model selection (hint-based + priority-based), auth enforcement,
budget checks, route restriction gates, and tag policy pre-auth
- Add elicitation_handler.py: MCP elicitation/create relay with
downstream client capability detection
- Wire sampling/elicitation callbacks in mcp_server_manager.py
gated behind allow_sampling/allow_elicitation config flags
- Add allow_sampling/allow_elicitation fields to MCPServer type
- Fix session lock deadlock: skip lock for JSON-RPC response POSTs
(elicitation/sampling replies) with truncated-body heuristic
- Extend client.py with sampling_callback and elicitation_callback
- Security: RouteChecks gate, tag-budget bypass fix, x-forwarded-for
spoofing fix, Latin-1 header encoding guard
- Add 4 new test modules (model access, priority selection, request
builder, tool conversion) + update existing MCP tests
* fix(security): run pre-call guardrails before MCP sampling acompletion
Without this, an upstream MCP server with allow_sampling enabled could
send prompts that bypass every guardrail (content filtering, PII
redaction, prompt-injection detection) configured on /chat/completions.
- Call proxy_logging_obj.pre_call_hook(call_type='acompletion') before
llm_router.acompletion so guardrails fire for sampling sub-calls
- Add HTTPException to the re-raise list so guardrail rejections
propagate correctly instead of being swallowed as generic errors
* feat(bedrock_mantle): add Responses API support (/openai/v1/responses) (#29490)
* feat(bedrock_mantle): add Responses API transformation config
* test(bedrock_mantle): cover trailing-slash api_base normalization
* feat(bedrock_mantle): export BedrockMantleResponsesAPIConfig
* feat(bedrock_mantle): register gpt-5.x Responses config (gpt-oss unchanged)
* feat(bedrock_mantle): add gpt-5.5/gpt-5.4 Responses price-map entries
* refactor(bedrock_mantle): exclude gpt-oss instead of allow-listing gpt-5 for Responses routing
Frontier OpenAI models on Bedrock Mantle are Responses-only on /openai/v1/responses;
gpt-oss is the legacy family that also speaks chat-completions. Gate by excluding
gpt-oss (which keeps its chat-completions emulation) and defaulting everything else
to the native Responses config, so future frontier models (gpt-6, etc.) route
correctly without a code change. Verified against the live us-east-2 Mantle endpoint:
gpt-oss 400s on /openai/v1/responses while gpt-5.5 400s on both standard paths.
* test(bedrock_mantle): cover supports_native_websocket opt-out
Closes the one uncovered line flagged by codecov on the Responses config.
The assertion documents that Mantle Responses has no realtime/websocket
transport, so realtime routing must not attempt a socket it cannot serve.
* fix(bedrock_mantle): route file_search through emulation instead of forwarding to Mantle
BedrockMantleResponsesAPIConfig inherited supports_native_file_search()
-> True from OpenAIResponsesAPIConfig but never overrode it. Mantle has no
OpenAI vector stores, so a forwarded file_search tool is rejected with a
400 (verified upstream: Tool type 'file_search' is not supported). Opting
out, like the existing supports_native_websocket override, routes the tool
through LiteLLM's file_search emulation instead.
* fix(bedrock_mantle): only route openai.gpt frontier models to Responses
The previous gate excluded gpt-oss and routed every other model to the
native Responses config. But on Mantle only the OpenAI gpt frontier models
(gpt-5.x) are served on /openai/v1/responses; gpt-oss and the non-OpenAI
families (nvidia, mistral, google, zai, ...) are chat-completions only and
400 on that path. Allow-list the openai.gpt- family (excluding gpt-oss)
instead, so chat-only models fall through to the chat-completions emulation.
Verified against the live us-east-2 endpoint: nvidia.nemotron-nano-9b-v2
returns 400 on /openai/v1/responses and 200 on /v1/chat/completions.
* feat(custom_llm): allow streaming/astreaming to yield ModelResponseStream (#27580)
* fix(custom_llm): allow streaming/astreaming to yield ModelResponseStream directly
* fix(streaming): enhance ModelResponseStream handling for custom LLM providers
* fix(streaming): strip finish_reason from content chunks and ensure tool_calls are preserved
* fix(streaming): add type ignore for finish_reason assignment in CustomStreamWrapper
* fix(proxy): strip stack trace from HTTP 503 responses (CWE-209) (#28330)
* fix(proxy/cwe-209): strip Python traceback from HTTP 503 error responses
The /cache/ping endpoint included a full Python traceback in its 503 error
response body (inside the ProxyException message), leaking internal file
paths, line numbers, and call stacks to any caller. Two MCP route handlers
in proxy_server.py similarly interpolated str(e) into "Internal server
error" detail strings.
Fix: log the traceback server-side via verbose_proxy_logger.exception()
and omit it from the ProxyException payload / HTTPException detail returned
to clients. Tests updated to assert no "traceback" keyword or frame paths
appear in the 503 body, with a new dedicated regression test.
CWE-209: Generation of Error Message Containing Sensitive Information.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(proxy/cwe-209): apply Greptile P2 fixes and add MCP exception-path tests
Greptile 4/5 review identified two remaining gaps and Codecov reported
0% coverage on the two MCP handler exception branches:
1. caching_routes.py — str(e) in "Service Unhealthy ({str(e)})" could
still leak Redis hostnames/IPs; replaced with static "Service Unhealthy".
HTTPException is now re-raised before the generic handler so the
"cache not initialized" 503 still reaches callers with its detail.
Removed the redundant str(e) arg from verbose_proxy_logger.exception()
(exception() already appends the traceback automatically).
2. tests — two new unit tests cover the exception paths in
dynamic_mcp_route and toolset_mcp_route that were previously at 0%:
- test_dynamic_mcp_route_unexpected_exception_returns_500_without_traceback
- test_toolset_mcp_route_unexpected_exception_returns_500_without_traceback
All 25 tests pass (9 caching + 16 MCP).
CWE-209: Generation of Error Message Containing Sensitive Information.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* test(caching_routes): restore precise assertion in test_cache_ping_no_cache_initialized
The assertion was weakened to `"Cache not initialized" in str(data)`, which
matches the raw string of the entire response dict and would pass even if the
error moved to an unexpected field or changed structure.
Restore a targeted check on the parsed response: assert the exact string in
the correct field `data["detail"]`, matching FastAPI's HTTPException
serialisation format {"detail": "<message>"}.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* test(caching_routes): restore precise assertion and add CWE-209 no-cache path test
The assertion in test_cache_ping_no_cache_initialized was weakened to
`"Cache not initialized" in str(data)`, which matched against the raw string
representation of the entire response dict. This would pass silently even if
the error message moved to an unexpected field or the structure changed.
Restore a targeted assertion on the parsed field:
assert data["detail"] == "Cache not initialized. litellm.cache is None"
matching FastAPI's HTTPException serialisation format exactly.
Add test_cache_ping_no_cache_does_not_expose_internals to show the code path
is still working correctly after the CWE-209 fix: verifies that the HTTPException
is re-raised as-is (no traceback, no source paths), and asserts the complete
response structure is exactly {"detail": "Cache not initialized. litellm.cache is None"}.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(caching_routes): restore ProxyException envelope for null-cache 503
The except HTTPException: raise guard (added in the CWE-209 fix) caused
the null-cache HTTPException to escape as FastAPI's {"detail": "..."} shape
instead of the {"error": {...}} ProxyException envelope that callers expect.
Move the null-cache guard before the try block and raise ProxyException
directly so the response structure is consistent with all other /cache/ping
503s, and the except HTTPException: raise guard is only reachable by
unexpected downstream HTTPExceptions.
Update the two no-cache tests to assert the correct ProxyException envelope.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* Update utils.py (#26609)
* feat(pricing): add Snowflake Cortex REST API model pricing (#26612)
* feat(pricing): add Snowflake Cortex REST API model pricing
## Summary
Adds pricing and context window information for 20+ Snowflake Cortex REST API models to `model_prices_and_context_window.json`.
## What's included
- **7 Claude models** (sonnet-4-5, sonnet-4-6, 4-sonnet, 4-opus, haiku-4-5, 3-7-sonnet, 3-5-sonnet) — with prompt caching rates
- **4 OpenAI models** (gpt-4.1, gpt-5, gpt-5-mini, gpt-5-nano) — with prompt caching rates
- **5 Llama models** (3.1-8b, 3.1-70b, 3.1-405b, 3.3-70b, 4-maverick)
- **1 DeepSeek model** (deepseek-r1)
- **1 Mistral model** (mistral-large2)
- **1 Snowflake model** (snowflake-llama-3.3-70b)
- **2 Embedding models** (arctic-embed-l-v2.0, arctic-embed-m-v2.0)
Each entry includes `input_cost_per_token`, `output_cost_per_token`, `cache_read_input_token_cost` (where applicable), `max_input_tokens`, `max_output_tokens`, and capability flags (`supports_function_calling`, `supports_vision`, `supports_prompt_caching`, `supports_reasoning`).
## Pricing source
All prices are in USD per token, sourced from the official [Snowflake Service Consumption Table](https://www.snowflake.com/legal-files/CreditConsumptionTable.pdf) — Tables 6(b) (REST API with Prompt Caching) and 6(c) (REST API).
## Context
The existing `snowflake/` provider has zero model entries in the pricing JSON, which means LiteLLM cannot track costs for Snowflake Cortex calls. This PR fills that gap.
## Related
- Existing provider: `litellm/llms/snowflake/`
- Cortex REST API docs: https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api
* Update model_prices_and_context_window.json
Fix the JSON parsing error
* Update model_prices_and_context_window.json
Removed the duplicate entry
* fix(utils): copy extra_body before adding unknown params to prevent model config mutation (#29620)
Fixes #29615. In add_provider_specific_params_to_optional_params, the line:
extra_body = passed_params.pop("extra_body", None) or {}
returns the original dict reference when extra_body is non-empty (truthy).
Subsequent writes like extra_body[k] = passed_params[k] then mutate the
shared model config object held by the router, poisoning /model/info and
all subsequent requests for that deployment.
The or {} short-circuit creates a new dict only when extra_body is falsy
(None or {}), which is why the bug does not reproduce with extra_body: {}.
Fix: wrap in dict() so we always work on a fresh shallow copy.
* fix(vertex_ai): Bake tool_choice into Gemini CachedContent body to prevent silent drop (#29097)
* fix(vertex_ai): bake tool_choice into Gemini CachedContent body to prevent silent drop
* address greptile feedback on tool_choice cache test
* adds test that uses ToolConfig(functionCallingConfig=FunctionCallingConfig(mode=ANY)) instead of a dict literal, mirroring what map_tool_choice_values actually produce
* fix(gemini/veo): move image from parameters into instances[0] (#29501)
* fix(gemini/veo): move image from parameters into instances[0]
Veo's predictLongRunning schema puts image (and prompt) on the
instances element; parameters is for aspectRatio/durationSeconds/etc.
The Gemini path was leaving image in params_copy, so it ended up
nested under parameters and the API silently ignored it.
The Vertex path already builds the instance dict explicitly, so this
just aligns the Gemini path with it.
Fixes #29498
* address greptile: unconditional pop + BytesIO test
- Pop `image` from params_copy unconditionally so it never reaches
GeminiVideoGenerationParameters even when None, removing implicit
reliance on Pydantic's extra-field-ignore.
- Add test_transform_video_create_request_image_filelike_goes_to_instance
covering the BytesIO path (_convert_image_to_gemini_format) — round-trips
the base64 to confirm encoding.
- Add test_transform_video_create_request_image_none_is_dropped covering
the new None branch.
* fix(huggingface): handle special token text in embedding usage (#29660)
* fix(guardrails): recompile ToolPermissionGuardrail rules on update_in_memory_litellm_params (#29655)
* fix(guardrails): recompile ToolPermissionGuardrail rules on update_in_memory_litellm_params
ToolPermissionGuardrail builds self.rules and the compiled target/pattern
maps only in __init__. The base update_in_memory_litellm_params re-sets raw
attributes via setattr but never rebuilds those maps, so a guardrail updated
in place (PUT /guardrails, or the immediate in-memory sync) keeps enforcing
the construction-time rules until it is reinitialized (PATCH path, periodic
DB poll, or restart).
Extract the compile step into _load_rules and override
update_in_memory_litellm_params to rebuild from it (dict- and model-safe),
re-normalizing default_action / on_disallowed_action. Mirrors the existing
PresidioGuardrail override of the same method. Adds regression tests.
Fixes #29592.
* fix(guardrails): handle dict params in ToolPermissionGuardrail in-memory update
Delegate to super() only for LitellmParams input (the base setattr loop is
model-only); apply the raw-dict case inline. Fixes the mypy arg-type error
and makes the recompile work when the proxy passes the raw DB dict.
* fix(guardrails): preserve tool-permission rules on a partial in-memory update
A partial update (e.g. a LitellmParams whose rules field is None) ran through
the generic setattr, which set self.rules to None, and the recompile was
skipped, leaving the guardrail with no rules. Snapshot the previous rules and
restore them when the update carries no rules; an explicit empty list still
clears them. Adds a regression test for the rules-absent case.
Addresses the Greptile review note on #29655.
* fix(bedrock): stop base_model label from stripping tools/tool_choice (#29621)
* fix(bedrock): stop base_model label from stripping tools/tool_choice
A Router/proxy Bedrock deployment whose model_info.base_model is a friendly
label (e.g. claude-haiku-4-5) silently lost tools/tool_choice: the outgoing
Converse request was built without toolConfig, so the model behaved as if no
tools were provided. Worked in v1.84.0, regressed in v1.85.0, and with
drop_params=true it failed silently.
Two changes compound into the bug. completion() passed model_info.base_model
as the model argument to get_optional_params, so the real Bedrock model id
never reached supported-param resolution; and get_supported_openai_params
resolved the provider config's params from base_model or model, letting the
label fully replace the real model. For Bedrock the label resolves to no tool
support, so tools/tool_choice were dropped before transformation.
completion() now keeps model as the real deployment model and threads the
resolved base_model (kwarg or model_info) through separately, and
get_supported_openai_params treats base_model as additive: it returns the
union of the params supported by model and by base_model. A hint can only add
capabilities, never strip ones the real model already exposes, which also
preserves the original base_model behavior from #27717 and Azure's base_model
driven model-type detection.
Fixes #29618
* test(main): make base_model param test robust to new parametrize cases
Restore an explicit per-case expected_model_param literal instead of
hardcoding the gemini id, so a future case with a different model can't
produce a misleading assertion failure.
* fix(fireworks_ai): pass response_format json_schema through unchanged (#29606)
FireworksAIConfig.map_openai_params was rewriting the OpenAI strict
`{type: json_schema, json_schema: {name, strict, schema}}` shape into
`{type: json_object, schema: ...}` before sending to Fireworks, dropping
`strict` and `name` and changing the `type`. Per Fireworks' docs json_object
means "force any valid JSON output (no specific schema)", so the schema
constraint was effectively dropped and grammar-guided decoding never ran;
model output silently violated the schema.
The rewrite landed in #7085 (Dec 2024) when Fireworks did not yet accept
native json_schema. Fireworks accepts the OpenAI strict shape natively now,
so the rewrite has become a regression.
Removes the rewrite. Passes response_format through unchanged. Updates the
existing test_map_response_format to assert pass-through. Adds focused
regression tests in tests/test_litellm/ covering preservation of type,
strict, name, and schema body, plus that json_object alone still works.
* fix(types): import Required from typing_extensions in gemini types
* style: reformat sampling_handler.py for py312 black compat
* refactor(mcp-sampling): extract helpers to fix PLR0915 too-many-statements in handle_sampling_create_message
* fix(proxy-server): add explicit ProxyLogging type annotation to proxy_logging_obj to fix mypy inference
* fix(mcp-sampling): suppress mypy assignment error on ImportError fallback for proxy_logging_obj
* fix(test): use .value when comparing LlmProviders enum against string in test_default_api_base
* fix(test): iterate LlmProviders enum in test_default_api_base to avoid str pollution from custom provider registration
litellm.provider_list is a mutable global initialized to list(LlmProviders) but custom_llm_setup() appends plain provider strings to it. When a test_custom_llm.py test runs first in the same xdist worker, provider_list contains a str and calling .value on it raises AttributeError. Iterate the immutable LlmProviders enum instead, which is deterministic and what the check intends.
* fix(mcp): depth-aware JSON-RPC response detection and neutral speed-priority fallback
Replace the flat substring check in the truncated-body routing path with a
top-level-key scan so a JSON-RPC response whose result payload nests a
"method" field is still detected as a response and skips the session lock,
removing a deadlock against the in-flight tool call awaiting it.
Drop the inverse max_output_tokens speed proxy when no model exposes
output_tokens_per_second; context-window size does not track latency, so a
neutral score avoids biasing speedPriority toward the smallest-context model.
* fix(guardrails): make ToolPermission rule reload atomic on invalid regex
_load_rules appended each rule to self.rules before compiling its regex, so an
invalid pattern raised mid-loop after the bad rule was already live but without
a _compiled_rule_targets entry. _matches_regex reads a missing compiled target
as a None pattern and returns True, turning the bad rule into a match-all that
silently applies its decision to every tool. Via update_in_memory_litellm_params
(PUT /guardrails) this corrupted the live guardrail.
Build the parsed rules and compiled maps into locals and swap them in only after
every regex compiles, and restore the previous ruleset if a live update is
rejected, so an invalid regex now fails the update without leaving the guardrail
enforcing a broken policy.
* test(mcp): cover sampling conversion, model resolution, and elicitation relay paths
The MCP sampling and elicitation handlers shipped with partial test
coverage, leaving the response-to-MCP conversion, the model resolution
fallback chain, completion-kwargs assembly, guardrail routing, and the
entire elicitation relay untested. That pulled the PR's diff (patch)
coverage below the codecov threshold even though overall project
coverage rose.
Add focused unit tests for _convert_openai_response_to_mcp_result,
_convert_mcp_tools_to_openai, _convert_mcp_tool_choice_to_openai, image
and audio content conversion, the hint-matching and fallback branches of
_resolve_model_from_preferences, _build_completion_kwargs, the router and
guardrail-rejection paths of _run_guardrails_and_call_llm, the
handle_sampling_create_message success and error-propagation flows, the
marker-hoisting fallback for tool content on unexpected roles, and the
elicitation form/url/generic relay together with its decline paths
---------
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: lengkejun <lengkejun@xd.com>
Co-authored-by: Yug <yugborana000@gmail.com>
Co-authored-by: Kent <72616338+kingdoooo@users.noreply.github.com>
Co-authored-by: tanmay958 <53569547+tanmay958@users.noreply.github.com>
Co-authored-by: DrishnaTrivedi <142084770+DrishnaTrivedi@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Navnit Shukla <Navnit.shukla25@gmail.com>
Co-authored-by: PRABHU KIRAN VANDRANKI <72809214+VANDRANKI@users.noreply.github.com>
Co-authored-by: Adrian Lopez <109683617+adriangomez24@users.noreply.github.com>
Co-authored-by: hcl <chenglunhu@gmail.com>
Co-authored-by: JooHo Lee <96564470+BWAAEEEK@users.noreply.github.com>
Co-authored-by: Dinesh Girbide <85330597+Dinesh-Girbide@users.noreply.github.com>
Co-authored-by: cloudwiz <22098246+andrey-dubnik@users.noreply.github.com>
Co-authored-by: Ahmad Khan <ahmadkhan2508@gmail.com>
Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com>
|
||
|
Sameer Kankute
|
ed073d382d
|
fix(gemini-realtime): use GA event names for Pipecat 1.3.x compatibility (#29662)
* fix(gemini-realtime): use GA event names for Pipecat 1.3.x compatibility Pipecat v1.3.0 adopted the OpenAI Realtime API GA event naming: response.audio.delta -> response.output_audio.delta response.text.delta -> response.output_text.delta response.audio.done -> response.output_audio.done response.text.done -> response.output_text.done The proxy was still emitting the old beta names; Pipecat's `parse_server_event` raises "Unimplemented server event type" for any unknown type, which killed the receive task handler and broke audio playback and tool-call delivery. Also: - conversation.item.created -> conversation.item.added (already handled) - client audio is buffered until backend setupComplete in deferred mode - call_id fallback UUID when Gemini returns empty id - status_details / token detail fields added to Pydantic-strict events The _GA_TO_BETA_EVENT_TYPES map in RealTimeStreaming already translates GA names back to beta for clients that opt in with the openai-beta header, so legacy clients are unaffected. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini-realtime): address greptile review comments - emit outputTranscription as response.output_audio_transcript.delta instead of suppressing it; GA_TO_BETA map handles translation for legacy clients - cap pre-setup audio buffer at 200 frames to prevent memory exhaustion; log a warning when the limit is hit and additional frames are dropped - log remaining dropped message count on flush error Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini-realtime): address veria review comments - remove unused OpenAIRealtimeConversationItemCreated import - fix guardrail bypass: semantic_vad early-return now preserves create_response when set so a guardrail-injected create_response:false is not silently dropped - add per-connection 10 MB byte cap alongside the 200-frame count cap for the pre-setup audio buffer to prevent memory exhaustion Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini-realtime): fix mypy arg-type on _finalize_gemini_live_setup setup parameter typed as BidiGenerateContentSetup to match the TypedDict passed at both call sites; was dict which mypy rejected. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini-realtime): widen _finalize_gemini_live_setup to Dict[str, Any] BidiGenerateContentSetup (TypedDict) is a subtype of Dict[str,Any] so both call sites (one passing a plain dict, one passing the TypedDict) satisfy mypy. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(gemini-realtime): cast BidiGenerateContentSetup to Dict at _finalize call site mypy rejects TypedDict as dict[str, Any] argument; cast at the call site where follow_up_setup is BidiGenerateContentSetup to satisfy the checker. Co-authored-by: Cursor <cursoragent@cursor.com> * Fix Gemini realtime beta compatibility * Fix deferred Gemini setup audio ordering * fix: preserve Gemini audio transcript ids * fix(realtime): cap pre-setup client buffer on all append paths Route every append to the deferred-setup pending buffer through the per-connection message/byte caps. Previously only the audio-buffer fast path enforced the caps; once one frame was buffered, a client that withheld session.update could stream arbitrary frames into _pending_messages_until_setup unbounded and exhaust proxy memory. * style(gemini-realtime): apply black formatting to transformation.py * fix(gemini-realtime): log beta-translation fallback and name native-audio marker Surface the previously swallowed exception in _send_event_to_client so a failed GA->beta translation is observable instead of silently forwarding the untranslated event. Extract the native-audio model substring used by _finalize_gemini_live_setup into a named constant documenting why speechConfig is dropped on those setups. --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Sameer Kankute
|
216c68db04
|
fix(gemini): googleSearch + server-side tools and googleMaps JSON schema (#29582)
* fix(gemini): keep googleSearch with server-side tools and googleMaps JSON schema Wire include_server_side_tool_invocations through completion() so mixed google_search and function tools are not dropped on Gemini 3+. Rewrite generationConfig to responseFormat when googleMaps is used with JSON schema. Fixes #27479 Fixes #29451 Co-authored-by: Cursor <cursoragent@cursor.com> * address greptile review feedback (greploop iteration 1) * style: fix black formatting in main.py for py312 compat * Fix Gemini Google Maps extra_body JSON rewrite --------- Co-authored-by: Cursor <cursoragent@cursor.com> |
||
|
Mateo Wang
|
cc55662e5f
|
fix(vertex): strip output_config.effort for Vertex Claude models that reject it (Haiku 4.5) (#29585)
* fix(vertex): strip output_config.effort for models that reject it Haiku 4.5 on Vertex AI does not support output_config.effort and 400s with "output_config.effort: Extra inputs are not permitted". PR #27074 emptied VERTEX_UNSUPPORTED_OUTPUT_CONFIG_KEYS so effort would forward for Opus/Sonnet 4.6+, but that made the strip unconditional across every Vertex Anthropic model, including ones that don't support it. Claude Code injects effort into its default Messages payload, so `claude --model claude-haiku-4.5` started failing. Make the sanitizer model-aware: drop output_config.effort for models that don't advertise output_config support (or any reasoning effort level) while forwarding it for those that do. The fix covers both the chat-completion and Messages pass-through transformation paths since they share the helper. * chore(vertex): log at debug when dropping unsupported output_config.effort Operators pointing an unregistered Vertex Claude alias that does support effort would otherwise see it stripped with no signal. Debug level keeps it out of normal logs since Claude Code sends effort on every request. |
||
|
Sameer Kankute
|
c7ab9adde5
|
Litellm oss staging 030626 (#29578)
* Fix incorrect agent API request example payload structure (#29556) * fix(otel): add litellm_metadata fallback in _get_span_context and _end_proxy_span_from_kwargs (#29427) * fix(otel): add litellm_metadata fallback in _get_span_context and _end_proxy_span_from_kwargs On /v1/messages and other LITELLM_METADATA_ROUTES, the parent OTel span is stored in litellm_params['litellm_metadata'] instead of litellm_params['metadata']. When the request body contains a native 'metadata' field (e.g. Anthropic's {"user_id": "..."}), litellm_params['metadata'] gets overwritten and the parent span is lost, producing orphan root spans with a different trace_id. Add fallback checks to litellm_metadata in: - _get_span_context(): so child spans find the correct parent - _end_proxy_span_from_kwargs(): so the proxy span gets closed Fixes: https://github.com/BerriAI/litellm/issues/27934 * test(otel): tighten assertions per Greptile review - test_span_context_metadata_takes_priority: assert litellm_metadata span is never accessed, proving metadata takes priority - test_span_context_no_parent_when_neither_has_span: assert both ctx and detected_span are None --------- Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Aneesh-Fiddler <aneeshfiddler@gmail.com> Co-authored-by: Sameer Kankute <sameer@berri.ai> * fix: remove premature end-user budget check from get_end_user_object (#29420) * fix(proxy): remove premature end-user budget check from get_end_user_object Problem: - `_check_end_user_budget()` was called inside `get_end_user_object()` - This caused budget checks to run BEFORE `skip_budget_checks` could be evaluated - Zero-cost models (e.g., local vLLM) were incorrectly blocked when end-users exceeded their budget, even though they should bypass budget checks Solution: - Remove `_check_end_user_budget()` calls from `get_end_user_object()` - Budget enforcement now happens exclusively in `common_checks()` where `skip_budget_checks` context is available - `get_end_user_object()` keeps `route` as optional in function parameter for backwards compatibility and future implementation. * refactor(tests): update budget enforcement tests to reflect changes in get_end_user_object - test_get_end_user_object() verifies data fetching - test_check_end_user_budget() verifies enforcement - test_budget_enforcement_blocks_over_budget_users() integrates _check_end_user_budget() - test_resolve_end_user_reraises_budget_exceeded() is now test_resolve_end_user since no budget exceeded is thrown in get_end_user_object() * Gemini /images/generate and /images/edits billing fixes + add support for size and aspect ratio params (#29534) * Fix Gemini image config mapping * Address Gemini image config review * Format Gemini image generation transform * Fix Gemini image token usage logging * Share Gemini image request helpers * Fix Gemini Imagen model routing * Fixes as per self code review * Fixes per internal code review * Stop gating Imagen imageSize forwarding * Document Gemini image size mapping source * chore: retrigger lint * Clarify Gemini candidate count precedence * Add Inception provider (#29522) * add inception as provider (chat, fim) * linting * seperate test suite for chat and fim * fix test coverage * fix: model hub custom pricing model info (#29293) * Opik user auth key metadata extractors (#28397) * fix: enhance Opik metadata extraction to include user API key auth context fixed after refactoring to extractor logic * test: add unit tests for OPik metadata extraction logic * fix: enhance extract_opik_metadata function to prioritize metadata sources for improved accuracy * fix(ci): clarified comments and edited unit tests * test: add unit tests for OPik metadata extraction with auth and requester overrides * fix(ui): replace fixed favicon.ico with current api get /get_favicon (#29532) Signed-off-by: José Luis Di Biase <josx@interorganic.com.ar> * fix(vertex/gemini): keep tool_call reference when a text-only assistant message follows (#29561) `_gemini_convert_messages_with_history` tracks `last_message_with_tool_calls` so a following tool result can be matched back to its tool call. The assignment was inside a branch guarded by `assistant_msg.get("tool_calls", []) is not None`, which is also True for a text-only assistant message (an empty list is not None). As a result, an assistant message with no tool calls that appears between a tool call and its tool result overwrote the reference, and conversion failed with: Exception: Missing corresponding tool call for tool response message. This shape is common: a model emits a short narration/assistant message after a tool call before the tool result is appended. Only update `last_message_with_tool_calls` when the assistant message actually carries tool_calls (or a function_call). Adds a regression test. Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> * Add 1-hour cache write pricing for EU/AU/JP Bedrock Anthropic models (#28572) * fix(thinking): handle None thinking param in is_thinking_enabled (#28598) Squash-merged by litellm-agent from Terrajlz's PR. * feat(helm): support tpl rendering in podAnnotations (#28609) Squash-merged by litellm-agent from devauxbr's PR. * Forward custom_llm_provider through the Responses API bridge (Fixes #28505) (#28575) * Forward custom_llm_provider through the Responses API bridge (Fixes #28505) When a Chat Completions request to a GPT-5.4+ model contains both `tools` and `reasoning_effort`, `completion()` auto-routes through `responses_api_bridge`. The bridge handler called `litellm.responses()` / `litellm.aresponses()` without forwarding the already-resolved `custom_llm_provider`, so the downstream call re-invoked `get_llm_provider()` with `custom_llm_provider=None` and stripped a second provider prefix from a `provider/provider/model` deployment string. For a deployment configured as `openai/openai/openai/gpt-5.5`, the bridge flow sent `openai/gpt-5.5` to the upstream API instead of the correct `openai/openai/gpt-5.5`. Upstream APIs that enforce model-name allow-lists rejected this as `key_model_access_denied`. Fix: pass the locally-resolved `custom_llm_provider` into both the sync `responses()` and async `aresponses()` calls so the downstream `_resolve_model_provider_for_responses` sees an explicit provider and skips the second prefix-strip. New regression test `tests/test_litellm/completion_extras/test_responses_bridge_provider_propagation.py` pins both call sites: each must forward `custom_llm_provider`. * fix(28505): set custom_llm_provider on request_data instead of as duplicate kwarg Greptile flagged that the previous patch passed custom_llm_provider as an explicit kwarg to responses()/aresponses() while request_data already carried it via the spread of sanitized_litellm_params, which would raise TypeError: got multiple values for keyword argument on every real bridge call. Switches to assigning request_data['custom_llm_provider'] before the call so the resolved provider wins over whatever sanitized_litellm_params spread in, without duplicating the kwarg. Updates the regression test to seed request_data with a sentinel custom_llm_provider so it actually exercises the overwrite path (the previous test mocked transform_request with a minimal dict and never hit the conflict). * chore: trigger shin-agent re-eval on retargeted staging base * chore: trigger shin-agent re-eval against updated Greptile state * Add 1-hour cache write pricing for EU/AU/JP Bedrock Anthropic models The 1-hour prompt-cache write tier (`cache_creation_input_token_cost_above_1hr`) was added to the us./global. variants of the Claude 4.5/4.6/4.7 family on Bedrock, but the eu./au./jp. cross-region inference profiles were left without it. AWS Bedrock pricing applies the same +10% regional premium across all geo profiles, so eu./au./jp. should carry the same 1-hour rates as us. (1.6x the 5-minute regional rate). Without these fields, cost tracking on EU/AU/JP Bedrock 1-hour-TTL prompt caching falls back to the 5-minute write rate and undercounts spend by ~60% for European, Australian, and Japanese tenants. Adds the 1-hour tier (and Sonnet 4.5's long-context >200K tier where AWS publishes one) to 14 regional Bedrock entries in both `model_prices_and_context_window.json` and the bundled `model_prices_and_context_window_backup.json`: - eu./au. Opus 4.6 ($11.00 / MTok) - eu./au. Opus 4.7 ($11.00 / MTok) - eu./au./jp. Sonnet 4.6 ($6.60 / MTok) - eu./au./jp. Sonnet 4.5 ($6.60 / MTok regular, $13.20 / MTok LC) - eu./au./jp. Haiku 4.5 ($2.20 / MTok) Also extends `tests/test_litellm/test_bedrock_anthropic_1hr_cache_pricing.py` with a `REGIONAL_EXPECTED` parametrized block covering all 13 new entries plus the existing 1.6x ratio invariant. Note: `eu.anthropic.claude-opus-4-5-20251101-v1:0` carries the wrong 5m rate today (base 6.25e-06 instead of regional 6.875e-06), which would break the 1.6x ratio check. It is intentionally left out of this PR so the scope stays "1-hour cache tier addition" — a separate follow-up should correct the EU 5m rates for Opus 4.5. --------- Co-authored-by: Terrajlz <info@jouleselectrictech.com> Co-authored-by: Bruno Devaux <devaux.br@gmail.com> Co-authored-by: Sameer Kankute <sameer@berri.ai> * Add 1-hour cache write pricing tier for Vertex AI Anthropic models (#28569) * fix(thinking): handle None thinking param in is_thinking_enabled (#28598) Squash-merged by litellm-agent from Terrajlz's PR. * feat(helm): support tpl rendering in podAnnotations (#28609) Squash-merged by litellm-agent from devauxbr's PR. * Forward custom_llm_provider through the Responses API bridge (Fixes #28505) (#28575) * Forward custom_llm_provider through the Responses API bridge (Fixes #28505) When a Chat Completions request to a GPT-5.4+ model contains both `tools` and `reasoning_effort`, `completion()` auto-routes through `responses_api_bridge`. The bridge handler called `litellm.responses()` / `litellm.aresponses()` without forwarding the already-resolved `custom_llm_provider`, so the downstream call re-invoked `get_llm_provider()` with `custom_llm_provider=None` and stripped a second provider prefix from a `provider/provider/model` deployment string. For a deployment configured as `openai/openai/openai/gpt-5.5`, the bridge flow sent `openai/gpt-5.5` to the upstream API instead of the correct `openai/openai/gpt-5.5`. Upstream APIs that enforce model-name allow-lists rejected this as `key_model_access_denied`. Fix: pass the locally-resolved `custom_llm_provider` into both the sync `responses()` and async `aresponses()` calls so the downstream `_resolve_model_provider_for_responses` sees an explicit provider and skips the second prefix-strip. New regression test `tests/test_litellm/completion_extras/test_responses_bridge_provider_propagation.py` pins both call sites: each must forward `custom_llm_provider`. * fix(28505): set custom_llm_provider on request_data instead of as duplicate kwarg Greptile flagged that the previous patch passed custom_llm_provider as an explicit kwarg to responses()/aresponses() while request_data already carried it via the spread of sanitized_litellm_params, which would raise TypeError: got multiple values for keyword argument on every real bridge call. Switches to assigning request_data['custom_llm_provider'] before the call so the resolved provider wins over whatever sanitized_litellm_params spread in, without duplicating the kwarg. Updates the regression test to seed request_data with a sentinel custom_llm_provider so it actually exercises the overwrite path (the previous test mocked transform_request with a minimal dict and never hit the conflict). * chore: trigger shin-agent re-eval on retargeted staging base * chore: trigger shin-agent re-eval against updated Greptile state * Add 1-hour cache write pricing tier for Vertex AI Anthropic models GCP Vertex AI publishes a separate 1-hour cache write column for the Claude family (1.6x the 5-minute write rate, matching the documented Bedrock ratio). LiteLLM's Vertex AI Anthropic entries only carry the 5-minute tier, so any request that uses `cache_control: {"ttl": "1h"}` on Vertex AI Claude is undercounted in cost tracking by ~60%. The runtime side already supports the 1-hour tier — `VertexAIAnthropicConfig` extends `AnthropicConfig`, populating `ephemeral_1h_input_tokens`, and `_calculate_cache_creation_cost` reads `cache_creation_input_token_cost_above_1hr`. Only the price registry was missing data. Adds the field to 19 vertex_ai/claude-* entries across both `model_prices_and_context_window.json` and the bundled `model_prices_and_context_window_backup.json`: - Haiku 4.5 ($1.25 -> $2.00 / MTok) - Sonnet 3.7 / 4 / 4.5 / 4.6 ($3.75 -> $6.00 / MTok) - Opus 4.5 / 4.6 / 4.7 ($6.25 -> $10.00 / MTok) - Opus 4 / 4.1 ($18.75 -> $30.00 / MTok) Adds `tests/test_litellm/test_vertex_anthropic_1hr_cache_pricing.py` mirroring the Bedrock equivalent — pins each (5m, 1h) pair per model and asserts the 1.6x ratio across the family. Fixes #27781. --------- Co-authored-by: Terrajlz <info@jouleselectrictech.com> Co-authored-by: Bruno Devaux <devaux.br@gmail.com> Co-authored-by: Sameer Kankute <sameer@berri.ai> * Fix Gemini multimodal function responses (#29325) Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> * address greptile review: add _transform_image_usage method and model-map supports_image_size flag - Add _transform_image_usage instance method to GoogleImageGenConfig that delegates to transform_gemini_image_usage, fixing the regression test - Replace hardcoded "2.5-flash" string check in supports_gemini_image_size with a get_model_info lookup on supports_image_size (default true) - Add supports_image_size: false to all gemini-2.5-flash model entries in model_prices_and_context_window.json so capability is controlled via the model map rather than embedded in code * fix test failures: schema validation, mypy type, model info plumbing, pricing test - Add supports_image_size to ModelInfoBase TypedDict so get_model_info surfaces it - Pass supports_image_size through _get_model_info_helper constructor call - Fix supports_gemini_image_size to use value is not False (None means unset, defaults to True) - Add supports_image_size to JSON schema in test_aaamodel_prices_and_context_window_json_is_valid - Correct gemini-3.1-flash-lite pricing assertions in test to match JSON values * Add Azure AI Kimi K2.6 metadata (#27052) * Add Azure AI Kimi K2.6 metadata * Scope Kimi metadata test cost map setup * fall back to substring check for models not in model_prices_and_context_window.json Models like gemini-2.5-flash-image-preview are not in the pricing JSON, so get_model_info raises. Fall back to "2.5-flash" not in model when the JSON has no explicit supports_image_size entry for the model. * fix(inception): don't forward global litellm.api_key to Inception FIM Match the Inception chat config: resolve only an Inception-specific key (param, litellm.inception_key, or INCEPTION_API_KEY) for the text-completion FIM path. The global litellm.api_key (often an OpenAI key) was both leaking to api.inceptionlabs.ai and taking precedence over the configured Inception key when set. * fix(auth): enforce end-user budget on custom-auth path that skips common_checks get_end_user_object() no longer raises BudgetExceededError, so custom-auth deployments with custom_auth_run_common_checks unset (which skip the centralized common_checks gate) stopped enforcing the end-user budget, letting an over-budget end user keep making requests. Re-enforce the budget in _run_post_custom_auth_checks on that path. --------- Signed-off-by: José Luis Di Biase <josx@interorganic.com.ar> Co-authored-by: Isha <72744901+IshaMeera@users.noreply.github.com> Co-authored-by: aneeshsangvikar <aneeshsangvikar@fiddler.ai> Co-authored-by: shin-berri <shin-laptop@berri.ai> Co-authored-by: yuneng-jiang <yuneng@berri.ai> Co-authored-by: Aneesh-Fiddler <aneeshfiddler@gmail.com> Co-authored-by: Suleiman Elkhoury <108065141+suleimanelkhoury@users.noreply.github.com> Co-authored-by: Dmitriy Alergant <93501479+DmitriyAlergant@users.noreply.github.com> Co-authored-by: Yanis Miraoui <yanis.miraoui19@imperial.ac.uk> Co-authored-by: Lovro Seder <vrovro@gmail.com> Co-authored-by: Thomas Mildner <12685945+Thomas-Mildner@users.noreply.github.com> Co-authored-by: José Luis Di Biase <josx@interorganic.com.ar> Co-authored-by: Lai Quang Huy <64073540+1qh@users.noreply.github.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com> Co-authored-by: Filippo Menghi <113345637+Cyberfilo@users.noreply.github.com> Co-authored-by: Terrajlz <info@jouleselectrictech.com> Co-authored-by: Bruno Devaux <devaux.br@gmail.com> Co-authored-by: ZHONG Ziwen <67355585+zzw-math@users.noreply.github.com> Co-authored-by: Emerson Gomes <emerson.gomes@thalesgroup.com> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
Mateo Wang
|
a5ccd96152
|
[internal copy of #29003] fix(vertex_ai): use user-supplied api_base as is for Model Garden OpenAI-compat path (#29530)
* fix(vertex_ai): use user-supplied api_base as is for Model Garden OpenAI-compat path
* chore(tests): url assertions and outputs
* fix(tests): fixing reference to unused test
* fix(aiohttp): drop octet-stream content-type on bodyless requests
The aiohttp transport forwarded httpx's empty request body straight
to aiohttp, which attaches a default Content-Type: application/octet-stream
for any bytes payload. Bodyless requests such as DELETE /responses/{id} then
hit OpenAI with that header and were rejected with unsupported_content_type,
breaking the e2e_openai_endpoints test_basic_response check. Coercing an
empty body to None makes aiohttp behave like the httpx transport and send no
content-type for bodyless requests.
---------
Co-authored-by: Steven Kessler <9701252+stvnksslr@users.noreply.github.com>
|
||
|
ryan-crabbe-berri
|
ebbc5cc787
|
feat(vector-stores): forward per-request params to Vertex AI Search (#29459)
* feat(vector-stores): forward per-request params to Vertex AI Search The vertex_ai/search_api search transform hardcoded the request body to query plus pageSize 10, dropping max_num_results and extra_body. Map max_num_results to pageSize and merge extra_body through with precedence, so callers can send native Discovery Engine fields such as dataStoreSpecs. Resolves LIT-3506 * fix(vector-stores): log effective query when extra_body overrides it When a caller passes a query inside extra_body, the outbound Vertex Search request used that value but model_call_details recorded the original, so the echoed search_query was stale. Log the effective query from the request body. * fix(vector-stores): allowlist Vertex AI Search extra_body fields Raw-merging extra_body let callers set dataStoreSpecs/branch to search a different Discovery Engine data store with the proxy's Vertex credentials, bypassing the vector_store_id path authorization. Reject target-selecting fields and forward only allowlisted per-request tuning fields. Resolves LIT-3506 * refactor(vector-stores): split Vertex AI Search extra_body allowlists by mode Data-store and engine/app serving configs accept different SearchRequest fields, so derive two TypedDicts (VertexSearchDataStoreExtraBody and VertexSearchEngineExtraBody) in types/vector_stores.py and make _filter_extra_body mode-aware via vertex_engine_id. dataStoreSpecs and numResultsPerDataStore now pass through in engine/app mode (where an app fans out across stores) and are rejected in data-store mode. branch/servingConfig/entity remain rejected in both modes. * fix(vector-stores): raise BadRequestError (400) for invalid Vertex Search extra_body Rejecting unsupported or target-selecting extra_body fields previously raised a bare ValueError, which the vector store error path mapped to a generic APIConnectionError (HTTP 500). Raise litellm.BadRequestError so invalid per-request input surfaces as HTTP 400 with a clear message. |
||
|
Sameer Kankute
|
e8fcb01215
|
Litellm OSS Staging (#29161)
* Cato Networks guardrail, based on Aim (#26597) * Aim was acquired by Cato Networks, creating Cato Networks guardrail based on Aim * Add more tests * Move test so they are reached by codecov coverage * base URL trailing slashes * Support Lemonade runtime context metadata (#28135) * Support Lemonade runtime context metadata * Add provider hook for runtime model metadata * Address provider model info review feedback Keep the runtime model info hook duck-typed instead of extending the base model-info class, and avoid importing ModelInfoBase from Ollama common utilities to reduce CodeQL cyclic-import noise. Co-authored-by: openhands <openhands@all-hands.dev> * Fix CI after staging rebase Relax the Ollama runtime metadata return annotation to match the provider-hook dict response and update the Google Interactions OpenAPI status expectation for the current live spec. Co-authored-by: openhands <openhands@all-hands.dev> * Normalize Lemonade runtime model metadata * Avoid leaking Ollama metadata auth * Avoid leaking Lemonade metadata auth --------- Co-authored-by: Graham Neubig <398875+neubig@users.noreply.github.com> Co-authored-by: openhands <openhands@all-hands.dev> * fix(cato): address guardrail review feedback Use proxy-authenticated user identity, forward moderation hook return values, and ensure streaming sender tasks are cancelled and awaited on exit. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(vertex_ai): route google/gemma-*-maas through partner-models OpenAI path - clone of #28010 (#28846) * fix(vertex_ai): route google/gemma-*-maas through partner-models OpenAI path Fixes #26083 vertex_ai/google/gemma-4-26b-a4b-it-maas previously fell through to the NON_GEMINI route. Per owtaylor's plan on #26083: add the google/gemma- prefix to PartnerModelPrefixes so is_vertex_partner_model picks it up and should_use_openai_handler routes it to the OpenAI-compatible /endpoints/openapi/chat/completions URL. No gemma-detection exclusion needed (the "gemma/" check uses a slash, which google/gemma-... doesn't match). No OpenAIGPTConfig subclass needed — works with the base handler. * fix(vertex_ai): mark gemma-4-26b-a4b-it-maas as vision-capable (empirically verified) * fix(vertex_ai): address greptile feedback — provider category, canonical URL, sync backup * test(vertex_ai): add function-calling and vision pass-through tests for Gemma MaaS Addresses oss-pr-review-agent-shin feedback on PR #28010: supports_function_calling, supports_tool_choice, and supports_vision were marked true but had no tests proving the payloads actually reached the OpenAI-compatible endpoint. Added: - test_gemma_maas_supports_function_calling — verifies the utility returns True when the model_cost entry carries supports_function_calling=true - test_gemma_maas_supports_vision — same for supports_vision - test_vertex_ai_gemma_function_calling_passthrough — verifies tools + tool_choice appear in the JSON body POSTed to /endpoints/openapi/chat/completions - test_vertex_ai_gemma_vision_passthrough — verifies image_url content parts survive transformation and reach the global endpoint URL * fix: Delete uv.lock * test(vertex_ai): add function-calling and vision pass-through tests for Gemma MaaS Addresses oss-pr-review-agent-shin feedback on PR #28010: P1 (patch target): Added a comment explaining why patching litellm.llms.custom_httpx.http_handler.AsyncHTTPHandler is correct — get_async_httpx_client() (defined in http_handler.py) instantiates AsyncHTTPHandler within that module's scope, so the definition-site patch intercepts it. Without the mock the test raises AuthenticationError, confirming it never silently passes. P2 (partner-provider regression guard): Added test_gemma_routes_through_openai_handler() which calls VertexAIPartnerModels.should_use_openai_handler() directly, so if Gemma's routing to VertexPartnerProvider.llama ever changes the URL-shape tests below it become a real regression guard rather than an unanchored unit test. Also added: - test_gemma_maas_supports_function_calling / supports_vision — capability flag checks via patch.dict(litellm.model_cost) - test_vertex_ai_gemma_function_calling_passthrough — tools + tool_choice forwarded in the request body - test_vertex_ai_gemma_vision_passthrough — image_url part survives transformation to the global endpoint Added: - test_gemma_maas_supports_function_calling — verifies the utility returns True when the model_cost entry carries supports_function_calling=true - test_gemma_maas_supports_vision — same for supports_vision - test_vertex_ai_gemma_function_calling_passthrough — verifies tools + tool_choice appear in the JSON body POSTed to /endpoints/openapi/chat/completions - test_vertex_ai_gemma_vision_passthrough — verifies image_url content parts survive transformation and reach the global endpoint URL * fix: proper patch for unit tests --------- Co-authored-by: Iana <iana@Shivakumars-MacBook-Pro.local> * fix(cato): guardrail all completion choices on output When n > 1, only choices[0] was analyzed and redacted. Iterate every Choices entry so block and anonymize actions apply to all completions. Co-authored-by: Cursor <cursoragent@cursor.com> * Fix review * fix(cato_networks): harden output anonymize handling and restructure nested UI routes Guard against empty redacted_output and empty all_redacted_messages from Cato. Restructure nested admin UI HTML exports to index.html so extensionless routes work. Co-authored-by: Cursor <cursoragent@cursor.com> * Fix mypy * fix(cato): guard missing policy_drill_down and all_redacted_messages keys * fix(cato): avoid KeyError bypassing block action on missing analysis_result * fix(cato): preserve non-text message fields during anonymize Rebuild redacted messages from the original messages, overwriting only content, so tool_calls, tool_call_id, name and multimodal fields survive the anonymize action. * fix(cato): preserve trailing messages when fewer redacted messages returned Avoid silently truncating the conversation in _anonymize_request when Cato returns fewer redacted messages than were sent, and isolate the no-api-key config test from a pre-existing CATO_API_KEY environment variable. * fix(cato,model-info): preserve stream block signal on sender teardown; forward api_key in dynamic model-info lookup Suppress ConnectionClosed (alongside CancelledError) when tearing down the Cato streaming sender task so a backend ConnectionClosed cannot mask the original StreamingCallbackError (e.g. a guardrail block) raised by the receive loop. Thread api_key through get_model_info -> _get_model_info_helper so an explicit key reaches a provider's dynamic get_model_info for a caller-supplied api_base. Previously only api_base was forwarded, so authenticated Ollama and Lemonade servers at a custom base could only be queried unauthenticated. * fix(cato): surface mid-stream forwarding errors instead of blocking on recv If the upstream LLM stream errors mid-flight, the sender task dies before sending the terminal done frame, so the consumer would block on websocket.recv() until Cato closes the connection. Race recv against the sender task and raise the stored sender exception promptly as a StreamingCallbackError. * fix(cato): drop spoofable end_user_id from guardrail user identity Only the key/JWT-bound user_email is a trusted identity. end_user_id is resolved from caller-supplied request fields (OpenAI user param, headers, metadata), so an authenticated caller with no bound user_email could set it to another user's email and have LiteLLM forward x-cato-user-email for that victim, poisoning Cato audit and policy attribution. Forward only user_email and omit the header otherwise. * fix(cato): harden output anonymize path against missing content key * fix(cato): fall back to original message when redacted content key is missing * refactor(model-info): drop unused api_key from cached model-info helper _cached_get_model_info_helper is only called by the cost-tracking hot path, which never authenticates, so the api_key parameter was never populated. Keeping it in the lru_cache key offered no benefit and risked fragmenting the high-RPS cache and retaining credential strings per entry. * fix(cato): preserve None content on tool-call-only choices in output hook * fix(ollama): respect static-model guard in OllamaConfig.get_model_info Delegate to OllamaModelInfo.get_model_info so statically-priced Ollama models short-circuit before the /api/show network call instead of hitting the server unconditionally. * fix(lemonade,ollama): treat empty api_key as unset to avoid leaking server creds An empty-string api_key was treated as an explicit key, so it passed the guard meant to keep server-side credentials off caller-supplied bases and then fell back through the env/global key chain. A caller could point api_base at a server they control and send api_key="" to receive the configured provider key in the Authorization header. Gate the credential fallback on the api_key being truthy instead of merely not-None. * fix(cato): inspect and redact Responses-API input, not just messages The guardrail only read data["messages"], so /v1/responses requests, which carry their text in data["input"], reached Cato as an empty message list and bypassed inspection entirely. Send build_inspection_messages(data) so both shapes are analyzed, and write anonymized results back with apply_redacted_messages_back when the request used input. * perf(utils): keep api_key out of get_model_info lru_cache key * fix(cato): propagate ssl_verify to streaming WebSocket connection The streaming hook applied ssl_verify only to the HTTP handler; the websockets.connect() call used default verification, so a custom Cato instance behind TLS with a self-signed cert worked for non-streaming calls but failed every streaming request. Resolve the ssl_verify setting into the connect() ssl argument, mirroring the HTTP handler. * refactor(utils): rename shadowing local in _get_model_info_helper * fix(cato): flatten multimodal chat content before inspection Chat Completions requests whose message content is a multimodal parts array were posted to Cato as the raw OpenAI parts, so text inside content: [{"type":"text", ...}] reached the model without Cato ever inspecting the string. Flatten each message's list content to plain text while keeping the list 1:1 with the request so the index-based redaction write-back stays valid; Responses-API input requests still go through build_inspection_messages. * test(lemonade): clear get_model_info cache around api_base test * fix(cato): inspect and redact Responses-API input even when messages present _inspection_messages returned early once messages was non-empty, so a /v1/responses caller could place benign text in messages and disallowed text in input and have only messages reach Cato while the model used input. Inspect both fields and write anonymize redactions back to input as well as the index-aligned messages. * test(log_db_metrics): assert table_name event_metadata contract log_db_metrics now emits minimal event_metadata via _safe_db_event_metadata (table_name only, function_name/function_kwargs/function_args dropped as redundant with call_type and unsafe to stamp on a span). The success-path test still asserted function_name membership and crashed with TypeError on the None metadata returned when no table_name is passed. Pass a table_name and assert the surfaced contract instead. * fix(cato): inspect and redact completion prompt and Responses-API instructions The Cato guardrail only inspected chat messages and the Responses-API input field, so blocked text placed in the legacy /v1/completions prompt or the /v1/responses instructions field reached the model without ever being sent to Cato. Both fields are now appended as synthetic inspection messages, and the anonymize path slices Cato's redactions back to the field they came from. * fix(cato): serialize non-str/bytes websocket chunks before forwarding * fix(cato): inspect tool descriptions and tool-call arguments * fix(cato): map redacted output by assistant index; restore get_model_info.cache_info * fix(cato): block output even when detection_message is null/empty A block_action returned by Cato on the output hook whose detection_message was null or empty was let through to the caller: the truthiness guard on detection_message skipped the HTTPException and the unblocked response was returned. Raise the HTTPException directly in _handle_block_action_on_output so the output path blocks unconditionally, mirroring the input path. * fix(cato): inspect and redact nested tool param and legacy function descriptions Tool/function parameter descriptions and the legacy functions[] array are forwarded to the model but were not seen by Cato, so blocked text hidden there bypassed inspection and anonymization. Recursively walk every description string in tools[].function and functions[] schemas for both the analyze payload and the anonymize write-back. * fix(cato): traverse schema descriptions iteratively to satisfy recursive detector The nested walk() generator recursed over tool/function JSON schemas with no depth bound, which the recursive_detector code-quality gate rejects. Replace it with an explicit-stack DFS that yields the same (container, key) refs in the same pre-order, so schema description redaction is unchanged. * fix(cato): inspect and redact response_format JSON schema descriptions response_format json_schema descriptions are forwarded to the model, so blocked text hidden in nested schema descriptions could bypass Cato inspection and redaction. Extend the schema-description walk to cover response_format alongside tools and legacy functions. * fix(cato): skip output rewrite when Cato returns no redaction Return None from call_cato_guardrail_on_output on monitor/no-action so the post-call hook only mutates the message when there is an actual redaction, instead of redundantly re-writing the original content. * refactor(utils): resolve explicit api_key model info without the cache Move the model-info build into a non-cached _build_model_info helper and drop api_key from the lru-cached _cached_get_model_info signature. Both cached helpers now take the same (model, provider, api_base) key and never forward api_key, while explicit per-caller keys are resolved through the builder directly instead of reaching into the cache wrapper's __wrapped__. * fix(cato): inspect and redact non-description schema string values Tool, function and response_format JSON schemas forward more than just description text to the model. enum, const, default, examples and title values are sent verbatim, so blocked content hidden in any of them bypassed Cato inspection and redaction. Walk those schema string values alongside descriptions on both the inspection and anonymize paths. * fix(model-info): surface swallowed dynamic model-info errors The provider-specific get_model_info dispatch falls back to the static cost map when a provider's dynamic lookup raises, which is intentional graceful degradation. Previously the exception was discarded with a bare debug line, so a real failure (e.g. a provider whose get_model_info signature does not accept api_key) was invisible. Log the exception at warning level with the model and provider context so the fallback is diagnosable. * fix(cato): inspect and redact Responses API output in post-call hook The post-call success hook only handled ModelResponse, so /v1/responses (which returns a ResponsesAPIResponse) bypassed the Cato output guardrail. Extract and inspect/redact every output_text content block and function-call arguments string, blocking on a block action, so generated text cannot escape inspection by using the Responses API. * chore: reset _experimental/out folder * chore(ui): remove orphaned prebuilt dashboard chunk files The _experimental/out manifests are byte-identical to the base branch, so the served dashboard already matches base. 436 unreferenced Next.js chunk files had accumulated in the directory and are not loaded by any manifest; removing them restores the committed UI artifacts to the base build and drops the artifact churn from this PR's diff. * fix(guardrails,ollama): forward ssl_verify to Cato init and raise_for_status on /api/show --------- Co-authored-by: Alex Yaroslavsky <trexinc@gmail.com> Co-authored-by: Graham Neubig <neubig@gmail.com> Co-authored-by: Graham Neubig <398875+neubig@users.noreply.github.com> Co-authored-by: openhands <openhands@all-hands.dev> Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: Piotr Placzko <piotr@icep-design.com> Co-authored-by: Iana <iana@Shivakumars-MacBook-Pro.local> Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com> |
||
|
ryan-crabbe-berri
|
1cce49b9d0
|
fix(vector-stores): support engines URL for Vertex AI Search (#27885)
Adds optional vertex_engine_id field to vertex_ai/search_api so users can route through a Discovery Engine search app instead of the data store directly. Required for website, healthcare, and connector-based data stores that return FAILED_PRECONDITION on the existing dataStores URL. Existing data-store-direct callers are unaffected. Resolves LIT-3036 |
||
|
Sameer Kankute
|
69afcd09d0
|
fix(vertex-ai): use DB credentials in video handlers + implement Veo video edit (#29098)
* fix(vertex-ai): pass litellm_params to validate_environment in video handlers and implement video edit for Veo - Pass litellm_params to validate_environment in 11 video handler call sites (remix, create_character, get_character, edit, extension, delete) so DB-stored Vertex AI credentials are used instead of falling back to ADC - Implement transform_video_edit_request/response for VertexAI: fetches source video via fetchPredictOperation then submits a new predictLongRunning request with the video bytes/gcsUri + edit prompt Co-authored-by: Cursor <cursoragent@cursor.com> * fix(vertex-ai): hoist fetchPredictOperation into handlers to avoid blocking event loop - Add get_video_edit_prefetch_params() to BaseVideoConfig (returns None) - VertexAI overrides it to return the fetchPredictOperation URL/body - Both sync and async video_edit handlers call this and use their shared httpx client for the fetch, passing the result as prefetched_source_data - transform_video_edit_request is now a pure transform with no HTTP calls - Fix extra_body.pop() mutation by working on a shallow copy Co-authored-by: Cursor <cursoragent@cursor.com> * fix(vertex-ai): include prefetch call inside _handle_error try/except block Co-authored-by: Cursor <cursoragent@cursor.com> * fix(videos): add prefetched_source_data param to all transform_video_edit_request overrides Co-authored-by: Cursor <cursoragent@cursor.com> * fix(video_edit): keep transform/pre_call outside try so validation errors propagate Move transform_video_edit_request and logging_obj.pre_call outside the try/except that wraps HTTP calls in (async_)video_edit_handler so that ValueError validation errors (e.g. 'source video not complete yet') are not silently wrapped as 500s by _handle_error. The prefetch HTTP call keeps its own try/except so its errors are still mapped through the provider's error handler. Matches the pattern used by video_extension_handler and video_remix_handler. Co-authored-by: Yassin Kortam <yassin@berri.ai> * refactor(vertex_ai): delegate get_video_edit_prefetch_params to status retrieve Co-authored-by: Yassin Kortam <yassin@berri.ai> * Fix varia review * fix(video_edit): route transform errors through _handle_error Wrap transform_video_edit_request and pre_call in the same try/except as the HTTP call in sync and async handlers so validation failures (e.g. source video not complete) return typed LiteLLM exceptions. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: Yassin Kortam <yassin@berri.ai> |
||
|
Sameer Kankute
|
3d0e0cee56
|
[Feat] Add tool calling support for gemini and vertex ai live api (#26590)
* Add tool calling support for gemini and vertex ai live api
* Fix greptile reviews
* Add new functionality behind flag
* fix greptile issues
* Fix greptile review
* Fix greptile review
* Fix greptile review
* Fix greptile review
* Fix greptile review
* fix lint
* fix(realtime): address P1 issues - guardrail timing and inputAudioTranscription default
- Remove early guardrail turn-detection update that consumed first setup slot
- Add inputAudioTranscription default in Gemini deferred-mode setup
- Add tests for both fixes
Made-with: Cursor
* fix(realtime): inject turn_detection into first session.update for deferred mode
- Instead of sending turn_detection as separate message (which gets dropped), inject it into the first client session.update
- This ensures guardrails work correctly in deferred mode
- Add test for turn_detection injection in deferred mode
Made-with: Cursor
* fix(realtime): emit response.created preamble before tool-call events
- Emit response.created, output_item.added, and conversation.item.created for function calls
- Ensures OpenAI Realtime API spec compliance
- Add test for preamble emission
Made-with: Cursor
* fix(realtime): add response.output_item.done to complete tool-call sequence
- Emit response.output_item.done between function_call_arguments.done and conversation.item.created
- Required by OpenAI Realtime spec to finalize function-call items
- Update test to verify complete event sequence
Made-with: Cursor
* fix(realtime): emit response.done after tool-call sequence (P0 CRITICAL)
- Add response.done event after tool-call loop to signal response completion
- Required by OpenAI SDK clients to submit tool results
- Without this, clients stall indefinitely waiting for response completion
- Update test to verify complete 6-event sequence including response.done
Made-with: Cursor
* fix(realtime): include function name in toolResponse (P1)
- Store call_id → name mapping when receiving toolCall from Gemini
- Look up and include name in functionResponses when sending tool results
- Required by Gemini Live API spec for proper tool call routing
- Add test to verify name field is included in round-trip
Made-with: Cursor
* fix: resolve merge conflict markers in UI build chunk
Take litellm_internal_staging version of e1a670efcb966aaa.js after
incomplete merge left conflict markers in the committed artifact.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(vertex_ai/realtime): call super().__init__() to initialize tool call state
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): correct guardrail flag and event-mapping fallback
- realtime_streaming: only mark _guardrail_turn_detection_update_sent
when the message was actually delivered to the backend. The provider
transformation (e.g. Gemini after initial setup) may silently drop
session.update; previously we set the flag anyway, falsely claiming
the disable was sent and preventing any retry on subsequent
session.created events. _send_to_backend now returns whether at
least one transformed message was sent.
- gemini realtime transformation: avoid shadowing the outer
openai_event variable in map_openai_event's fallback loop. With
the new toolCall entry now last in MAP_GEMINI_FIELD_TO_OPENAI_EVENT,
an unmatched key would otherwise leak FUNCTION_CALL_ARGUMENTS_DONE
and skip the ValueError raise. Use a distinct loop variable so the
is-None check correctly raises for unknown Gemini messages.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini/realtime): reset response IDs after tool-call response.done
After closing a tool-call response, clear current_output_item_id and
current_response_id so post-tool model turns emit a fresh response.created
preamble. Add regression tests and align guardrail turn_detection test with
GA session shape; apply Black formatting.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix lint
* fix(realtime): log injected message and forward guardrail VAD-disable on Gemini
- Move store_input() after the guardrail turn_detection injection in
client_ack_messages so audit logs reflect what is actually forwarded
to the backend (previously the unmodified pre-injection message was
logged).
- In Gemini's _handle_session_update, allow a session.update that only
carries a turn_detection change to be forwarded as a follow-up Gemini
setup with realtimeInputConfig.automaticActivityDetection set, even
after the initial setup. This restores the guardrail layer's ability
to disable VAD auto-response in non-deferred mode (the default Gemini
flow), which was a regression after _handle_session_update started
silently dropping subsequent session.update messages. Both flat
beta-style and nested GA-style turn_detection payloads are accepted.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini/realtime): resolve mypy TypedDict errors in transformation
Align realtime event payloads and setup types with OpenAI/Gemini TypedDicts so mypy passes and tool-call events type-check correctly.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(realtime): forward turn_detection updates for Vertex; respect partial VAD config; cache setup after send
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): consolidate send-and-cache, guard session.update lookup, preserve client turn_detection in GA remap
- Replace duplicated transform/send/cache logic in client_ack_messages with a call to _send_to_backend so future changes stay in one place.
- VertexAIRealtimeConfig.transform_realtime_request now uses .get('session') or {} for the first session.update so a malformed client payload no longer crashes the connection.
- Move the audio-transcription guardrail turn_detection injection to run BEFORE the beta->GA session remap. This lets the injected create_response ride along with any client-provided turn_detection fields (e.g. silence_duration_ms) into the nested audio.input.turn_detection path produced by the remap instead of being stranded as a separate root-level dict.
- Update the deferred-mode injection test to assert the GA-shaped location.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): pop tool_call_id mapping after use to bound memory
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): correct deferred-setup session.created modalities and reset IDs after response.done
- Convert provider's real session.created to session.updated when a synthetic
one was already forwarded so clients receive the authoritative modalities
derived from their session.update instead of the synthetic placeholder.
- Reset current_response_id / current_output_item_id after Gemini RESPONSE_DONE
so a toolCall arriving in a later frame starts a fresh response instead of
reusing the completed response's ID and emitting a duplicate response.done.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini-realtime): preserve nested turn_detection through map_openai_params
After the GA remap moves session.turn_detection into session.audio.input.turn_detection,
Gemini's map_openai_params only looks at top-level keys and silently drops it. Normalize
the extracted turn_detection back to the top level on first session.update so the guardrail
create_response:False (and any client-provided VAD settings) reach the Gemini setup.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): normalize Vertex AI nested turn_detection and unify session.created guardrail ordering
- Vertex AI _build_vertex_ai_setup_config now lifts nested
audio.input.turn_detection to the top level before calling
map_openai_params, mirroring the parent GeminiRealtimeConfig
behavior. Without this, guardrail-injected create_response: False
was silently dropped for GA-protocol Vertex AI clients.
- realtime_streaming session.created handling now sends the
(possibly re-typed) event first and then triggers the guardrail
turn-detection update for both first and duplicate cases, removing
the inconsistent guardrail-then-event ordering for duplicates.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): tolerate non-dict turn_detection in guardrail injection
When a client sends a session.update whose turn_detection field is None or
a non-dict value (e.g. "auto"), the guardrail injection used setdefault
followed by item assignment on the returned value, raising TypeError. The
inner except only caught JSONDecodeError/AttributeError, so the TypeError
escaped to the outer Exception handler that wraps the entire client_ack
loop, killing the connection. Replace non-dict turn_detection with a
fresh dict carrying create_response=False so the guardrail still applies
without crashing the loop.
* fix(gemini realtime): default synthetic session.created modalities to AUDIO
The synthetic session.created event emitted in deferred setup mode used
TEXT as the default for responseModalities, while _handle_session_update
defaults to AUDIO. Align the default so clients reading modalities from
the initial session.created see the correct value for live sessions.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(vertex_ai/realtime): drop follow-up session.update to avoid 1007 close
Vertex AI Live treats setup as a first-and-only client message; emitting a
second setup with realtimeInputConfig only closes the websocket with a 1007
policy error. Reverting the follow-up-setup branch restores the pre-existing
no-op behavior for subsequent session.update messages.
* fix(gemini realtime): default responseModalities to AUDIO in delta events
Align return_new_content_delta_events with the AUDIO defaults used in
_handle_session_update and transform_session_created_event so deferred
session config does not produce TEXT-typed delta events for audio data.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): default response.done modalities to AUDIO and correct audio-done test
* fix(realtime): set guardrail turn_detection flag only after successful send
Previously the _guardrail_turn_detection_update_sent flag was set inline
during message rewriting in client_ack_messages, before the modified
session.update was forwarded to the backend. If _send_to_backend raised
(e.g. backend WebSocket disconnect), the exception was caught and the
loop continued, but the flag remained True — permanently disabling the
guardrail create_response=False injection for the rest of the session.
Neither the client_ack_messages path nor the
_maybe_send_guardrail_turn_detection_update backup path would retry.
Track the injection locally and only set the flag after _send_to_backend
returns a truthy sent result, matching the pattern used by
_maybe_send_guardrail_turn_detection_update.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(vertex_ai realtime): keep VAD enabled when guardrails inject create_response: False
map_automatic_turn_detection sets disabled=True whenever create_response is
absent OR False. Transcription guardrails inject create_response: False to
suppress auto-responses while expecting VAD to stay active, but the previous
override in _build_vertex_ai_setup_config only fired when create_response was
absent, leaving disabled=True and silently breaking speech detection and
transcription events. Vertex Live has no 'VAD on, no auto-response' mode, so
always keep VAD active in the setup config.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): normalize GA-remapped session fields before mapping
map_openai_params only recognises the flat OpenAI-beta keys (modalities,
input_audio_transcription, turn_detection). For GA clients the upstream
shim renames these into the nested GA schema (output_modalities,
audio.input.transcription, audio.input.turn_detection), causing them to
be silently dropped in _handle_session_update. Add a normalization helper
that surfaces the GA-remapped values back at the top level so the
existing mapping logic picks them up. Without this, a GA client
explicitly requesting modalities=['text'] would still default to audio
output.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(vertex_ai/realtime): normalize all GA-remapped session fields before mapping
Previously _build_vertex_ai_setup_config only lifted nested turn_detection
back to the top level. GA clients' output_modalities and
audio.input.transcription were silently dropped because map_openai_params
only recognises the flat OpenAI-beta keys. Use the parent's
_normalize_session_payload_for_mapping so modalities, transcription, and
turn_detection are all surfaced before mapping.
* fix(realtime): force create_response=False in all client session.update turn_detection when audio guardrails active
Prevents a client from re-enabling Gemini/GA VAD auto-response (and thereby
bypassing the audio transcription guardrail) by sending a later
session.update with turn_detection.create_response: true.
* fix(lint): silence PLR0915 on client_ack_messages
The function exceeded the 50-statement limit (64 > 50) after recent
realtime guardrail additions. Matches the existing project pattern for
inherently complex event/message-mapping methods (see _process_event,
translate_messages_to_responses_input, transform_realtime_response,
_arealtime, etc.).
* fix(gemini realtime): preserve original setup config on follow-up session.update
Gemini Live treats a second BidiGenerateContentSetup as a full session
replacement, not a partial merge. The guardrail-driven turn_detection-only
session.update was emitting a setup containing only model + realtimeInputConfig,
which would silently drop tools, generationConfig, inputAudioTranscription, and
systemInstruction from the original setup. Carry forward the cached original
setup and only override realtimeInputConfig.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): avoid double-serialization and normalize non-dict turn_detection in guardrail override
- Skip the force-override block when the injection block already ran for
the same session.update to avoid redundant JSON re-serialization.
- Normalize non-dict client-provided turn_detection values (flat and
nested audio.input.turn_detection) to a dict before enforcing
create_response=False, matching the injection block's behavior and
preventing potential bypass on backends that accept non-dict values.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* test(gemini realtime): exercise toolCall → function_call_output name round-trip
Update test_gemini_realtime_function_call_output_transformation to pre-load
the call_id → name mapping by transforming a Gemini toolCall first, then
assert that the resulting Gemini toolResponse functionResponses entry
carries the function name. This pins the production round-trip rather
than the degenerate 'name missing' branch.
* fix(realtime): correct conversation_id, VAD disable, modality state, empty toolCall
- Gemini tool-call response.done now includes conversation_id so clients
can match it against the preceding response.created.
- Vertex AI setup no longer overrides an explicit guardrail-injected
create_response: False back to disabled: False; the guardrail's intent
to disable VAD auto-response is now respected.
- Modality handler is now passed the locally-updated response/item IDs
rather than the original input snapshot, preventing stale IDs after a
prior tool-call/response.done in the same JSON message resets them.
- Skip emitting orphaned response.created/response.done events when
Gemini sends an empty functionCalls array.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): preserve client session.update fields on follow-up Gemini setup
In non-deferred mode the auto-setup pre-populates session_configuration_request,
so a later client session.update carrying tools or instructions used to fall
into the subsequent path and only forward turn_detection. Rebuild a merged
follow-up setup that overlays the new client fields on top of the original
setup so tools/instructions/etc. are no longer silently dropped.
* fix(gemini realtime): include usage on tool-call response.done; coerce non-dict tool output to struct
- Tool-call response.done now includes an empty usage object, matching the
non-tool-call path so OpenAI-compatible clients always see usage.
- _handle_function_call_output wraps non-dict JSON parses under a 'result'
key so Gemini's functionResponses[].response (a Struct) always receives a
mapping.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): deep-merge nested config in follow-up session update
Previously, the follow-up setup performed a shallow merge between the
original setup and new overrides. If a session.update touched any field
inside generationConfig (e.g. modalities), the entire generationConfig
would be replaced, silently dropping unrelated sub-keys like temperature
or maxOutputTokens. Apply the same deep-merge to realtimeInputConfig so
partial automatic-activity-detection updates don't drop other realtime
input config fields either.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): default conversation_id before tool-call response.done
mypy flagged that response.done's conversation_id (str on the TypedDict)
could be None when current_response_id was already set on entry. Ensure
the fallback runs unconditionally before the response is constructed.
* fix(realtime): deep-merge generationConfig and refresh cache on follow-up setup
A subsequent Gemini session.update that touches any generationConfig sub-field
(e.g. just temperature) was clobbering the original generationConfig — silently
dropping responseModalities and switching the session to text-only. Deep-merge
generationConfig so existing keys (responseModalities, maxOutputTokens, ...) are
preserved when the client updates only a subset.
Also drop the early-return in _cache_session_configuration_request so the
cached payload tracks the latest setup sent to the backend. Without this,
downstream readers (transform_session_created_event, modality lookup in
return_new_content_delta_events) keep reading stale modalities/system
instruction after a follow-up setup.
* fix(gemini realtime): mirror modalities/temperature/max_output_tokens on tool-call response.created
The audio/text response.created preamble includes modalities, temperature,
and max_output_tokens on the response object so spec-compliant clients can
initialise per-response state. The tool-call response.created was missing
these fields, leaving clients without consistent response metadata when a
response starts with a tool call instead of content. Read them from the
cached session_configuration_request the same way the audio/text path
does.
* fix(gemini realtime): keep call_id→name mapping across function_call_output retries
A client SDK that retries function_call_output (or sends the same result
twice) would previously hit a missing-name lookup on the second send
because _handle_function_call_output popped the call_id → name entry.
Without name, Gemini may silently reject the response. Use dict.get so
the mapping persists for the lifetime of the session.
* fix(gemini realtime): empty toolCall must not terminate the WebSocket
If Gemini sends a toolCall whose functionCalls list is empty (or absent),
the previous `continue` left returned_message empty and the
"Unknown message type" guard fired, killing the WebSocket session.
Return a normal (empty) result instead so the session keeps going.
* fix(vertex realtime): warn when dropping guardrail turn-detection update
In non-deferred mode the auto-setup is sent on connect, so the audio-transcription
guardrail's subsequent session.update carrying turn_detection.create_response=False
cannot be forwarded as a second setup (Vertex Live closes the WebSocket with 1007).
Surface a warning when this specific drop happens so operators know the model
will auto-respond before the guardrail can gate it, instead of failing silently
at debug level.
* fix(gemini realtime): deep-merge automaticActivityDetection on follow-up session.update
The follow-up setup merge already deep-merged generationConfig and
realtimeInputConfig, but realtimeInputConfig.automaticActivityDetection
itself is a nested dict. A partial VAD update (e.g. the
guardrail-injected disabled=True from create_response=False) silently
dropped unrelated knobs such as silenceDurationMs and prefixPaddingMs
from the original setup. Deep-merge that block too so partial overrides
only touch the fields they specify.
* fix(realtime): record synthetic session.created in deferred-setup mode
The deferred-setup path emits a synthetic session.created directly to
the client websocket but did not run it through RealTimeStreaming's
store_message, so the event was missing from the session log used by
success_handler / async_success_handler. Call store_message before
forwarding so the synthetic event lands in the same log stream as
provider-driven events.
* fix(gemini realtime): bound _tool_call_id_to_name with an LRU; exercise modality forwarding test
Two minor follow-ups from review:
* Switch _tool_call_id_to_name to a 256-entry LRU OrderedDict so a long
session with many tool calls doesn't grow the dict without bound,
while retried function_call_output lookups still hit for recently-seen
call_ids.
* Fix test_gemini_realtime_transformation_session_created to wrap the
cached session config in {"setup": ...} so the modality lookup in
transform_session_created_event actually exercises responseModalities
forwarding (the prior payload was silently treated as empty).
* test(gemini realtime): wrap remaining cached session configs in setup envelope
The session_configuration_request the proxy caches is always serialized
as {"setup": ...}; three modality-related tests dumped a bare config
dict instead, so transform_session_created_event's
`.get('setup', {})` quietly returned an empty dict and the
responseModalities lookup ran against the default rather than the
fixture. Wrap the remaining tests in the same shape the production
cache uses so any regression in modality forwarding actually trips.
* fix(gemini realtime): cast merged realtimeInputConfig for typeddict assignment
mypy flagged the assignment of the merged dict into
BidiGenerateContentSetup.realtimeInputConfig with [typeddict-item]: the
intermediate variable widens to dict[Any, Any], losing the TypedDict
narrowing the previous dict-literal form had.
* test(gemini realtime): wrap test_gemini_tool_call_resets_ids fixture in setup envelope
The cached session_configuration_request the proxy stores is always
serialized as {"setup": ...}; this test passed a bare config dict, so
transform_session_created_event's .get('setup', {}) returned an empty
dict and the responseModalities lookup ran against the default rather
than the fixture. Wrap the fixture in the same shape the production
cache uses.
* fix(gemini realtime): skip unknown sibling keys in transform loop
Gemini realtime messages can include sibling metadata keys like
usageMetadata alongside primary payload keys (toolCall, serverContent).
Previously, the transform loop called map_openai_event for every
top-level key, raising ValueError for unknown ones and terminating
the WebSocket session.
Skip top-level keys not present in MAP_GEMINI_FIELD_TO_OPENAI_EVENT
to keep the session alive when Gemini emits usage metadata with a
toolCall response.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): scope dotted-key event lookup and propagate session metadata to tool-call response.done
- map_openai_event: only check the current key/value pair when resolving
dotted map entries (e.g. serverContent.turnComplete) so a sibling key in
the same frame can't misclassify the event being processed
(e.g. toolCall returning RESPONSE_DONE).
- tool-call path: extract generationConfig once and include modalities,
temperature, and max_output_tokens on response.done so its shape matches
response.created and the non-tool-call response.done.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): cast maxOutputTokens to int for typeddict assignment
* fix(gemini realtime): use camelCase maxOutputTokens in response.done
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): cast maxOutputTokens to int for typeddict assignment
* fix(realtime): inject guardrail turn_detection on subsequent session.update without one
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): tolerate sibling-only frames (e.g. standalone usageMetadata)
A Gemini Live frame that contains only metadata keys outside
_KNOWN_GEMINI_TOP_LEVEL_KEYS (e.g. a bare {"usageMetadata": {...}}
emitted between turns) leaves returned_message empty after the
transform loop and was tripping the 'Unknown message type' guard,
which raised ValueError and terminated the WebSocket session.
Treat such frames as no-ops and return the unchanged state instead.
* fix(gemini realtime): preserve sibling toolCall when serverContent has only transcription
Previously, when a Gemini frame contained both a transcription-only
serverContent and a sibling toolCall, the transcription handler would
early-return and silently drop the toolCall. Instead, mark serverContent
as handled and fall through so the main loop still processes siblings
like toolCall, while preserving the prior no-op behavior for empty/
transcription-only frames.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* refactor(gemini realtime): drop unused json_message arg from map_openai_event
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): promote nested turn_detection when flat value is not a dict
When the session payload had `turn_detection: None` (or any non-dict value), the
normalizer skipped promoting the GA nested `audio.input.turn_detection` because
it only checked key presence. The stale None then flowed into
`map_automatic_turn_detection` and raised TypeError on `'create_response' in value`.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(realtime): run guardrails on function_call_output content
Tool result outputs are client-controlled and fed to the model, so
they must pass the same content checks as user text messages.
Otherwise an attacker can smuggle blocked content into a
function_call_output and have the model process it.
* fix(gemini realtime): emit function_call_arguments.delta before .done
Gemini delivers the full function-call arguments in a single toolCall
frame. The OpenAI Realtime spec orders the streaming events as
output_item.added -> function_call_arguments.delta(+) ->
function_call_arguments.done -> output_item.done. Emit a single delta
carrying the complete arguments string before the matching .done so
spec-compliant SDK clients that accumulate deltas and gate finalisation
on at least one delta arriving do not stall on Gemini tool calls.
* fix(realtime): avoid stale session.created flag triggering guardrail re-injection
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(ci): restore guardrail injection on duplicate session.created and cast realtime delta event
- Re-enable the one-time guardrail turn_detection update on duplicate
session.created. `_maybe_send_guardrail_turn_detection_update` is
already idempotent via `_guardrail_turn_detection_update_sent`, so
the previous guard was unnecessary and broke the deferred-setup path
where the synthetic session.created is emitted by llm_http_handler
outside this loop (no prior chance to inject).
- Cast the response.function_call_arguments.delta dict appended to
`returned_message: List[OpenAIRealtimeEvents]` so mypy is satisfied.
* fix(realtime): forward sanitized function_call_output on guardrail block
Providers that pair every toolCall with a toolResponse (e.g. Gemini and
Vertex Live) stay in the awaiting-tool-call state until a toolResponse
arrives. Dropping a blocked function_call_output outright left those
providers stalled — the subsequent guardrail clientContent and
response.create were ignored because the prior toolCall had no matching
toolResponse.
When the client-supplied tool output fails the realtime guardrail check,
forward a sanitized placeholder function_call_output (same call_id,
generic policy marker as output) instead of dropping the message
entirely. The placeholder carries no blocked content, so the model never
sees it, while still completing the provider's tool-call cycle so the
session can recover and the violation message reaches the user.
* fix(gemini realtime): preserve sibling keys on empty toolCall no-op
Replace the early return on `functionCalls` empty/absent with a
`continue` plus a `tool_call_handled` flag that mirrors the existing
`server_content_handled` pattern. The post-loop guard already
distinguishes intentionally-consumed known keys from genuinely-unknown
messages, so adding `toolCall` to that exclusion list lets the loop
continue iterating over any sibling top-level keys in the same Gemini
frame instead of short-circuiting on the first empty toolCall.
In practice Gemini's protobuf places `toolCall`/`serverContent`/
`setupComplete` in a `oneof` so the only realistic sibling is
`usageMetadata` (already filtered as unknown-top-level), but the
uniform handling avoids silently discarding any future sibling key
should the wire format grow.
* fix(gemini realtime): redact realtime payloads from debug logs
The transform_realtime_response debug logs were dumping the raw inbound
Gemini frame and each outbound OpenAI event payload (up to 500 chars).
Realtime frames carry transcripts, model output, and tool-call arguments,
so those strings ended up in application logs whenever DEBUG was enabled.
Replace the inbound dump with just the top-level frame keys and the
outbound dump with just the event type.
* fix(realtime): check function_call_output before user role to prevent guardrail bypass
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): propagate usageMetadata on tool-call response.done
Gemini Live emits usageMetadata as a sibling top-level key alongside the
toolCall frame; the tool-call branch was unconditionally building
response.done from get_empty_usage(), so tokens consumed by tool-call
turns were recorded as zero spend and bypassed LiteLLM budget
accounting. Mirror the non-tool-call RESPONSE_DONE path: when the same
frame carries usageMetadata, run VertexGeminiConfig._calculate_usage and
forward the real token counts.
* fix(realtime): send sanitized toolResponse before guardrail clientContent
Two related fixes for the function_call_output blocked-by-guardrail path:
1. Ordering: Gemini Live requires a matching toolResponse immediately
after a toolCall before any other client message. Previously we ran
the guardrail first (which sends clientContent/cancel) and only then
forwarded the sanitized function_call_output. Add an optional
pre_block_backend_message arg to run_realtime_guardrails so the
sanitized toolResponse is emitted before the guardrail's own backend
messages.
2. Stale pending flag: stop setting _pending_guardrail_message in the
tool-output block. That flag exists to swallow the reflexive
response.create an OpenAI client sends right after a user text
message. In tool-calling flows the client may never send a
response.create (e.g. Gemini SDKs auto-respond), so leaving the flag
set would consume an unrelated response.create from a later turn.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* test(model_prices): allow audio_transcription_config in schema
* fix(gemini realtime): event_id, item copy, and dict guard for tool-call events
- Emit event_id on response.output_item.added for tool calls so spec-compliant
OpenAI Realtime SDK clients can index/deduplicate the event like every other
server-sent event in the sequence.
- Pass a shallow copy of function_call_item to response.output_item.done and
conversation.item.created so downstream handlers (e.g. the beta-protocol
translator) that mutate the item dict don't corrupt sibling events sharing
the same reference.
- Guard map_openai_event against non-dict values (e.g. Gemini's
'setupComplete: true' boolean payload) so the WebSocket session doesn't die
with an AttributeError on the unguarded .get() call.
Add NotRequired event_id field on OpenAIRealtimeStreamResponseOutputItemAdded
to keep existing call-sites that don't set event_id compatible.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(gemini realtime): buffer standalone usageMetadata for next response.done
Gemini Live can emit usageMetadata as a standalone WebSocket frame between
turns. The previous transformer treated those frames as no-ops, so token
counts arriving outside the closing turnComplete/toolCall frame were
dropped from spend and budget accounting. An authenticated client could
drive turns whose usage was recorded as zero, bypassing budgets.
Buffer any standalone usageMetadata on the config instance and attribute
the deferred counts to the next emitted response.done (tool-call or
normal). In-frame usageMetadata remains authoritative and clears the
buffer.
* merge main (#28839)
* fix(helm): drop main- prefix from default image tag (#28710)
* fix(helm): drop main- prefix from default image tag
The default image tag in the deployment + migrations-job templates was
`main-{{ .Chart.AppVersion }}`. The current release pipeline publishes
content tags without the `main-` prefix (e.g. `v1.85.1` / `1.85.1`,
`v1.86.0-rc.1` / `1.86.0-rc.1`), so the rendered ref points at a tag
that does not exist on GHCR or DockerHub and installs fail with
ImagePullBackOff.
- templates/deployment.yaml, templates/migrations-job.yaml: render
`.Chart.AppVersion` directly instead of `main-<AppVersion>`.
- Chart.yaml: bump stale `appVersion: v1.80.12` (not on either
registry) to `v1.85.1` so local-checkout installs also resolve.
- values.yaml: update the commented tag-override hint to match.
* fix(helm): use :latest in tag override example, not pinned version
Per review: ghcr.io/berriai/litellm-database:latest is a floating
alias for the most recent stable (same digest as :main-stable),
maintained by the release pipeline's UPDATE_LATEST advance step.
Better example than a pinned version that goes stale.
* test(model_prices): allow audio_transcription_config in schema (#28708)
The schema in test_aaamodel_prices_and_context_window_json_is_valid uses
additionalProperties: false. The azure/speech/azure-stt entry added in
#27482 introduced an audio_transcription_config field that the schema
did not whitelist, so the test fails on every branch built on top of
staging.
Add the field as a string property.
* fix(team): refresh team cache on team_model_add/delete (LIT-3244) (#28683)
* fix(team): refresh team cache on team_model_add/delete (LIT-3244)
team_model_add and team_model_delete wrote to the DB but did not
invalidate the in-memory LiteLLM_TeamTableCachedObj used by
common_checks. After the v1.83.14 common_checks centralization made
team.models authoritative on /v1/files and /v1/vector_stores/*,
adding a Team-BYOK model silently failed to grant the new public
model name to team members until the cache TTL expired (and a
removed model kept working until then on the symmetric path).
Extract the cache-refresh snippet from update_team into a small
helper and apply it consistently at all three team-write sites.
* test: also assert updated models in team-cache-refresh pin
Strengthens the LIT-3244 regression test to also assert
`call_kwargs["team_table"].models` matches the updated row,
not just `team_id`. Both `existing_team` and `updated_team`
share `team_id` in the test setup, so the previous assertion
would have passed even if the implementation accidentally cached
the pre-mutation row.
Greptile review feedback.
* fix(team): hydrate object_permission on cache-refreshing team updates
The Prisma update calls in update_team, team_model_add, and
team_model_delete returned a team row with object_permission_id set
but object_permission=None (the relation was not requested via
include=). _refresh_cached_team then wrote that to the in-memory
LiteLLM_TeamTableCachedObj, and the cache-hit path in get_team_object
returns the cached object without re-hydrating. Downstream consumers
(validate_key_search_tools_against_team, the MCP/agent authz paths)
treat a missing object_permission as no team-level restriction, so
a team-write op silently dropped object-permission enforcement until
the cache TTL expired or a DB-fetch path re-hydrated it.
Add include={"object_permission": True} to all three updates so the
refresh writes a complete cached team. Extend the LIT-3244 regression
test to pin both the cached object_permission and the include shape
on the Prisma call.
Surfaced in PR review of LIT-3244.
* fix(ui/add-model): stop vertex_ai-anthropic_models from leaking under Anthropic (#28723)
`getProviderModels()` matched a model into a provider's dropdown when the
model's `litellm_provider` string *contained* the provider key as a
substring. The intent was to admit suffix variants (e.g. `anthropic_text`,
`bedrock_converse`), but the substring check is too loose: it also pulls in
unrelated providers whose name happens to contain the key, most visibly
`vertex_ai-anthropic_models` matching `anthropic` and `vertex_ai-openai_models`
matching `openai`.
Replace `.includes()` with separator-anchored prefix matching
(`startsWith(provider + "_")` / `startsWith(provider + "-")`). All legitimate
variants in `model_prices_and_context_window.json` still match
(`anthropic_text`, `azure_text`, `azure_ai`, `bedrock_converse`,
`bedrock_mantle`, `cohere_chat`, `fireworks_ai-embedding-models`,
`vertex_ai-*`, `vertex_ai_beta`), and the cross-provider leak is closed.
Tests: update one assertion that pinned the buggy substring behavior
(`custom_openai_endpoint` matching `openai` — not a real provider value);
add 6 new tests covering the leak regressions and the variant-preservation
contract for vertex_ai/bedrock/fireworks.
* Fix spend logs v2 route permissions (#28705)
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: ryan-crabbe-berri <ryan-crabbe-berri@users.noreply.github.com>
* fix(proxy): Bedrock Knowledge Base pass-through: preserve SigV4 headers and signed request body (#27526)
* Fix Bedrock KB pass-through SigV4 headers and signed body
Coerce botocore HeadersDict to a dict for pass-through routes. When
forward_headers is true, drop request headers that collide case-insensitively
with signed headers so client Bearer auth does not shadow AWS SigV4.
Send prepped.body as raw content so the outbound payload matches the
signature after logging hooks mutate the parsed dict.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Simplify pass-through raw body handling
Read the SigV4-signed bytes directly from request.state inside
pass_through_request instead of threading a custom_raw_body argument
through three functions. Helper methods are restored to their original
signatures, and the new branch lives in one place at each httpx call site.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Harden pass-through raw body read from request.state
Guard missing request.state (test fixtures) and ignore non-bytes/str
values so MagicMock does not trigger the SigV4 raw-body path.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Test pass_through_request state_raw_body uses httpx content=
Cover non-streaming (async_client.request) and streaming (build_request)
paths so SigV4 bytes on request.state are not replaced by json= of a
hook-mutated dict.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* chore(tests): migrate Bedrock CI to AWS account 941277531214 (#28728)
* chore(tests): migrate Bedrock CI from AWS account 888602223428 to 941277531214
The original account (888602223428) was put under a security restriction by
AWS after a root access key leaked in a PR comment. While that account works
its way through the AWS Support unlock process, Bedrock-touching CI tests have
been migrated to a fresh account (941277531214).
Changes:
- Replace 26 hardcoded references to 888602223428 with 941277531214 across
8 files (provisioned-model ARNs, imported-model ARNs, AgentCore runtime
ARNs, batch execution role ARN, and example proxy config).
- The provisioned-model and imported-model ARNs are referenced only from
mocked unit tests — no AWS resources to recreate.
- The batch execution IAM role has been recreated in the new account with
the same name and equivalent permissions.
- The two AgentCore runtimes (hosted_agent_r9jvp-3ySZuRHjLC,
hosted_agent_13sf6-cALnp38iZD) are being recreated in the new account
under the same names — see tools/agentcore-deploy/ in a follow-up.
CircleCI env vars AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY / AWS_REGION_NAME
were updated separately via the CircleCI API to point at the new account.
Smoke-tested locally against the new account:
aws bedrock-runtime converse --region us-west-2 \
--model-id us.anthropic.claude-sonnet-4-5-20250929-v1:0 \
--messages '[{"role":"user","content":[{"text":"ping"}]}]'
→ 200, model returned 'pong'
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* chore(tests): refresh AgentCore ARN suffixes to match newly-deployed runtimes
The first migration commit replaced just the account ID, but AgentCore
auto-assigns a random 10-char suffix to every runtime on creation — we
can't reuse the original suffixes (`3ySZuRHjLC`, `cALnp38iZD`) in the
new account. Updated the AgentCore-runtime ARNs in the three files that
reference real runtime IDs (not the mock-based unit-test ARNs).
Deployed runtimes:
arn:aws:bedrock-agentcore:us-west-2:941277531214:runtime/hosted_agent_r9jvp-Rq79QFC2fp
arn:aws:bedrock-agentcore:us-west-2:941277531214:runtime/hosted_agent_13sf6-4046UzHSwy
Both runtimes are status=READY and pass a smoke invoke:
$ aws bedrock-agentcore invoke-agent-runtime --agent-runtime-arn ... --payload '{"prompt":"ping"}'
→ 200, {"result": "echo: ping"}
The agent is a minimal echo (see /tmp/agentcore_deploy/agent.py for the
deploy artifacts). Tests that only verify the SDK wiring will pass; if any
test asserts on agent output content, swap the echo for the real agent.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* chore(tests): point Bedrock batch tests at new-account S3 bucket
The account migration (888602223428 -> 941277531214) was a flat
account-ID swap, which only rewrites ARNs that embed the account
number. S3 bucket names carry no account ID, so the live Bedrock
batch tests still uploaded to `litellm-proxy` — a bucket that lives
in the old account. S3 names are globally unique, and the old account
still holds that name, so it can't be recreated in the new account.
Rename to `litellm-proxy-941277531214` (account-ID suffix guarantees
global uniqueness). The bucket must be created in 941277531214 and the
batch execution role granted s3:GetObject/PutObject/ListBucket on it
before this job is run in CI.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* chore(tests): point live S3 logging test at new-account bucket
Same account-ID-free blind spot as the batch bucket: `load-testing-oct`
lives in the old account and its name can't be reused globally. The
`logging_testing` CI job is wired into the workflow and runs
test_basic_s3_logging, which uploads to this bucket with the CI env
creds, then lists and deletes objects — a live dependency.
Rename to `load-testing-oct-941277531214`. The bucket must exist in the
new account with the CI IAM principal granted
s3:PutObject/GetObject/ListBucket/DeleteObject before this job runs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* chore(tests): repoint Bedrock guardrail IDs to new-account guardrails
The migration left guardrail IDs untouched (no account ID in them), so
all live guardrail tests failed with "guardrail identifier or version
does not exist" against 941277531214. Recreated both guardrails in the
new account and updated the hardcoded IDs:
- wf0hkdb5x07f -> zgkmukebruil (PII mask: PHONE + CREDIT_DEBIT_CARD,
with explicit inputAction=ANONYMIZE so masking applies to INPUT,
which is the source litellm's moderation hook sends)
- ff6ujrregl1q -> 4w3d1di3snt5 (blocks "coffee"; blocked message set
to the exact string the tests assert on)
Updated test_bedrock_guardrails.py, otel_test_config.yaml, and the
guardrailConfig in test_bedrock_completion.py. Verified locally: the 5
previously-failing guardrail tests now pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(bedrock): migrate legacy models to current inference profiles
The new CI account (941277531214) cannot invoke legacy Bedrock models
(AWS gates them: "marked by provider as Legacy... not actively using in
the last 30 days"). Migrated the live-call tests:
- anthropic.claude-3-sonnet-20240229 -> us.anthropic.claude-sonnet-4-5-20250929-v1:0
- anthropic.claude-3-haiku-20240307 -> us.anthropic.claude-haiku-4-5-20251001-v1:0
Current Claude models on Bedrock require the us. inference-profile prefix
(bare on-demand ids are rejected).
cohere.command-r-plus has no working replacement (all Cohere is legacy-
gated in the new account): swapped to claude-haiku-4-5 in provider-
agnostic param lists. amazon.titan-image-generator skipped (no working
replacement). Mocked/transformation/cost tests that reference the legacy
strings are intentionally left unchanged. Verified live against the new
account.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(bedrock): repoint SageMaker + Knowledge Base to new-account resources
These referenced account-scoped resources by hardcoded id that only
existed in the old account, so the migration's account-ID swap missed
them. Recreated in 941277531214 and repointed:
- SageMaker endpoint jumpstart-dft-hf-textgeneration1-mp-20240815-185614
-> litellm-ci-textgen (gpt2 on a TGI container, ml.g5.xlarge)
- Bedrock Knowledge Base T37J8R4WTM -> LCYXFBR2TU (OpenSearch Serverless
vector store + titan-embed-text-v2, seeded with a LiteLLM doc)
Verified live: test_sagemaker.py (12 passed) and
test_bedrock_knowledgebase_hook.py (12 passed).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(reasoning_effort_grid): skip bedrock claude-opus-4-7 cells (not entitled on 941277531214)
claude-opus-4-7 is listed in the new Bedrock CI account's foundation
models but invoke is denied (AccessDeniedException: "not available for
this account"). Bedrock access to the flagship Opus requires an AWS
Sales request, not the self-serve model-access toggle, so it can't be
enabled inline with the rest of the account migration.
Add an optional `skip_reason` to ModelEntry and set it on the
bedrock-claude-opus-4-7 entry; the grid test honors it via pytest.skip.
Cell count (231) and route coverage are unchanged, so the structural
asserts still pass. Restore coverage by deleting the one skip_reason
line once access is granted.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* test(bedrock): swap/skip legacy-gated models unavailable on new CI account
The migrated AWS account (941277531214) cannot access several models that
the old account could, so the remaining red CI jobs were hitting real
Bedrock "Access denied / Legacy" and "account not authorized" errors:
- image_gen: skip both Nova Canvas test classes (amazon.nova-canvas-v1:0 is
legacy-gated), matching the existing titan skip.
- batches: skip test_async_file_and_batch (Bedrock batch inference is not
authorized on the new account; requires an AWS support case).
- litellm_overhead: swap legacy claude-3-5-haiku for the active
us.anthropic.claude-haiku-4-5 inference profile.
- test_completion_claude_3_function_call: swap legacy claude-3-sonnet for the
active us.anthropic.claude-sonnet-4-5 inference profile.
https://claude.ai/code/session_01Y7zgHYu9GX29YRwV4yiWAa
* test(bedrock): fix remaining e2e legacy-model + batch failures on new CI account
- e2e_openai_endpoints: skip test_bedrock_batches_api (Bedrock batch inference
is not authorized on account 941277531214) and migrate the missed
s3_bucket_name in oai_misc_config.yaml to litellm-proxy-941277531214.
- build_and_test: swap legacy bedrock claude-3-sonnet for the active
us.anthropic.claude-sonnet-4-5 inference profile in the proxy structured
output e2e test.
https://claude.ai/code/session_01Y7zgHYu9GX29YRwV4yiWAa
* test(bedrock): make opus-4-7 + batch cells fail loudly and mock image-gen (#28791)
Replace the silent skips added for the new CI account with noisier behavior:
- reasoning-effort grid: opus-4-7 cells now fail (when AWS creds are present)
instead of skipping, so the missing entitlement stays visible in CI; they
still skip when AWS creds are absent (local dev)
- Bedrock batch inference tests: drop the skip so they run and fail until
batch access is granted
- Titan + Nova Canvas image-gen tests: mock the Bedrock HTTP call so the
transform + cost-tracking path stays under test without live model access
https://claude.ai/code/session_01MT7SWDnXUjv6e6EPG7BDjT
Co-authored-by: Claude <noreply@anthropic.com>
* test(bedrock): use pytest.xfail for known-failing opus-4-7 cells
Replace pytest.fail with pytest.xfail when a model has a fail_reason,
so known-broken cells stay visible as XFAIL without keeping CI red.
Co-authored-by: Yassin Kortam <yassin@berri.ai>
---------
Co-authored-by: Mateo <mateo@Mateos-MacBook-Pro.local>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(otel): export SERVER span on management-endpoint success without http_request (#28794)
Co-authored-by: Yassin Kortam <yassinkortam@Yassins-MacBook-Pro.local>
* chore(ci): merge dev branch (#28801)
* chore(proxy): route path-dependent call sites through get_request_route
Replace direct ``request.url.path`` reads in auth, ACL, routing, and
audit-log decisions with ``get_request_route(request)`` — the helper
already added in ``auth/auth_utils.py`` that returns the ASGI
``scope["path"]`` with ``root_path`` stripped. Starlette reconstructs
``url.path`` from the Host header; ``scope["path"]`` is uvicorn's
parse of the request line and matches what FastAPI dispatches on, so
it's the authoritative route for any decision that should agree with
the actual handler.
Sites:
- _experimental/mcp_server/auth/user_api_key_auth_mcp.py
- management_endpoints/mcp_management_endpoints.py
- vector_store_endpoints/utils.py
- pass_through_endpoints/pass_through_endpoints.py
- auth/route_checks.py
- litellm_pre_call_utils.py
- spend_tracking/spend_management_endpoints.py
- common_utils/http_parsing_utils.py
- management_helpers/utils.py
- health_endpoints/_health_endpoints.py
Adds regression tests in tests/proxy_unit_tests/test_proxy_routes.py
that construct a Request with scope["path"] set to a benign route and
the Host header crafted so url.path would resolve differently; each
site's decision is asserted against scope["path"].
* chore(proxy): make get_request_route imports lazy at call sites
Move the ``from litellm.proxy.auth.auth_utils import get_request_route``
imports added in the prior commit back to the function bodies that use
them. The module-level form participates in a long-standing import
cycle through ``auth_utils -> _types -> ...`` and was flagged by CodeQL
on the PR; the lazy form matches the pattern the proxy already uses
for ``user_api_key_auth`` and related helpers elsewhere in these files.
Also drop the ``RouteChecks._is_assistants_api_request`` delegation in
``_get_metadata_variable_name`` introduced in the prior commit — the
delegation pulled ``RouteChecks`` into the same cycle, and the call
site reuses the resolved route for its other branches, so inlining
the substring check is both cycle-free and avoids a redundant second
``get_request_route`` call.
Comment in test_proxy_routes.py acknowledges that the two MCP table
entries exercise ``get_request_route`` directly rather than the full
production handler (which needs ASGI scope + MCP state to invoke).
---------
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: user <70670632+stuxf@users.noreply.github.com>
* chore(ci): merge dev branch (#28657)
* feat(dashboard): navbar hierarchy + Agent Platform notifications (#27543)
* feat(dashboard): refine navbar zones and Agent Platform notice
Restructure the admin navbar for production users: clear product vs community
vs personal columns with vertical dividers, icon-only Slack/GitHub in a
shared chip, and Docs/Blog typography aligned on an 8px rhythm.
Add a notifications bell with popover linking to the LiteLLM Agent Platform
repo and optional mark-as-read persistence.
Promote the account control with initials avatar, single-line display name,
and navDisplayName mapping for placeholder user ids (e.g. default_user_id).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(dashboard): address PR review — AntD buttons, public page guard, dedupe regex
- Replace raw <button> with AntD Button in BlogDropdown, NotificationsBell, UserDropdown, and test mock
- Guard NotificationsBell + container behind !isPublicPage to avoid rendering on public pages
- Remove redundant equality checks in navDisplayName (regex already covers them)
- Remove unused `lower` variable after simplification
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
* fix(dashboard): drop dead useHealthReadiness import in navbar
The module was removed in #27896 (replaced by useHealthReadinessDetails),
but the import survived the rebase. The symbol is unused — only
useHealthReadinessDetails is consumed in the file. Removing the dead
import unblocks the UI TypeScript build.
* fix(dashboard): align CommunityEngagementButtons test with icon-only aria-labels
The component was refactored to an icon-only chip with aria-label='LiteLLM
on GitHub' (squash #27543), but the test still asserted /star us on
github/i. Update the query to match the rendered accessible name.
* refactor(dashboard): drop unused props from NavbarProps
The navbar refactor moved user identity + dark-mode state to internal
hooks (useAuthorized, useWorker), but the NavbarProps interface still
declared userID, userEmail, userRole, premiumUser, isDarkMode, and
toggleDarkMode as required, forcing every caller to thread them through.
Drop them from the interface and all four call sites (page.tsx,
(dashboard)/layout.tsx, public_model_hub.tsx, navbar.test.tsx). Also
shrinks the destructure in layout.tsx so the now-unused locals stop
being pulled out of useAuthorized().
* refactor(dashboard): use useSyncExternalStore for NotificationsBell dismiss flag
Reads/writes of the litellmHideAgentPlatformBanner key were done
directly inside NotificationsBell via a useEffect + useState pair.
Every other localStorage-backed flag in the dashboard (Disable
ShowPrompts, DisableBouncingIcon, DisableShowNewBadge,
DisableUsageIndicator, DisableBlogPosts) is wrapped in a
useSyncExternalStore hook over localStorageUtils so all mounted
components stay in sync.
Extract useHideAgentPlatformBanner to follow the same shape, swap
NotificationsBell to consume it, and add a regression test that
two sibling bells stay in sync without a remount when one is
dismissed.
* refactor: mask credential fields in proxy settings GET responses (#28682)
* refactor: mask credential fields in proxy settings GET responses
Brings SSO settings, cache settings, and the email/Slack alerting view in
/get/config/callbacks in line with the HashiCorp Vault config-override
pattern, so persisted credentials are not transported back to the UI in
plaintext.
* refactor: harden short-value masking and hoist alerting var constant
Closes two review observations:
- mask_sensitive_keys now replaces short values (below the visible
prefix+suffix length) with an all-mask string instead of returning them
unchanged, so a 1-7 character credential is no longer round-tripped
verbatim.
- _ALERTING_SENSITIVE_VARS is moved out of get_config() to a module-level
constant, matching the analogous _SSO_SENSITIVE_FIELDS and
_CACHE_SENSITIVE_FIELDS in the SSO and cache endpoint files.
---------
Co-authored-by: Krrish Dholakia <krrish+github@berri.ai>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix(ui): show 2-decimal precision for max_budget on key overview (#28809)
The Key Info Overview tab's Spend card truncated sub-dollar budgets to
"$0" because formatNumberWithCommas defaults to 0 decimals. The Settings
tab passes 2; align the overview so a $0.10 budget renders as "$0.10".
Resolves LIT-2845
* feat(proxy): allow `llm_api_routes` virtual keys to list MCP servers (#28442)
* feat(proxy): allow llm_api_routes virtual keys to list MCP servers
Add a new `mcp_discovery_routes` group (GET /v1/mcp/server and GET
/v1/mcp/server/{server_id}) and include it in `llm_api_routes` so that
virtual keys configured with `allowed_routes=["llm_api_routes"]` can
discover the MCP servers they have access to. Previously these calls
failed with 'Virtual key is not allowed to call this route. Only allowed
to call routes: [llm_api_routes]'.
The GET handlers already sanitize the response for restricted virtual
keys via `_sanitize_mcp_server_list_for_virtual_key`, stripping
credential-bearing fields (url, headers, env). Write methods
(POST/PUT/DELETE) on the same paths remain gated by the existing
handler-level admin role checks.
The new discovery list is intentionally kept OUT of
`mcp_inference_routes`, so `is_llm_api_route()` still returns False
for these paths — this preserves the existing contract that
DISABLE_LLM_API_ENDPOINTS must not block the Admin UI from listing MCP
servers.
Co-authored-by: ryan-crabbe-berri <ryan-crabbe-berri@users.noreply.github.com>
* refactor(proxy): make MCP discovery carve-out method-aware
Replace the `mcp_discovery_routes` group in `llm_api_routes` with a
method-aware special case inside `is_virtual_key_allowed_to_call_route`.
Virtual keys with allowed_routes=["llm_api_routes"] are now permitted
to call only GET /v1/mcp/server and GET /v1/mcp/server/{server_id} —
non-GET methods and multi-segment admin sub-paths fall through to the
existing 403. This keeps the general llm_api_routes list free of
management paths and avoids accidentally exposing POST/PUT/DELETE
writes through the route-check layer.
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: ryan-crabbe-berri <ryan-crabbe-berri@users.noreply.github.com>
* chore(ci): merge dev branch (#28807)
* chore(proxy): route path-dependent call sites through get_request_route
Replace direct ``request.url.path`` reads in auth, ACL, routing, and
audit-log decisions with ``get_request_route(request)`` — the helper
already added in ``auth/auth_utils.py`` that returns the ASGI
``scope["path"]`` with ``root_path`` stripped. Starlette reconstructs
``url.path`` from the Host header; ``scope["path"]`` is uvicorn's
parse of the request line and matches what FastAPI dispatches on, so
it's the authoritative route for any decision that should agree with
the actual handler.
Sites:
- _experimental/mcp_server/auth/user_api_key_auth_mcp.py
- management_endpoints/mcp_management_endpoints.py
- vector_store_endpoints/utils.py
- pass_through_endpoints/pass_through_endpoints.py
- auth/route_checks.py
- litellm_pre_call_utils.py
- spend_tracking/spend_management_endpoints.py
- common_utils/http_parsing_utils.py
- management_helpers/utils.py
- health_endpoints/_health_endpoints.py
Adds regression tests in tests/proxy_unit_tests/test_proxy_routes.py
that construct a Request with scope["path"] set to a benign route and
the Host header crafted so url.path would resolve differently; each
site's decision is asserted against scope["path"].
* chore(proxy): make get_request_route imports lazy at call sites
Move the ``from litellm.proxy.auth.auth_utils import get_request_route``
imports added in the prior commit back to the function bodies that use
them. The module-level form participates in a long-standing import
cycle through ``auth_utils -> _types -> ...`` and was flagged by CodeQL
on the PR; the lazy form matches the pattern the proxy already uses
for ``user_api_key_auth`` and related helpers elsewhere in these files.
Also drop the ``RouteChecks._is_assistants_api_request`` delegation in
``_get_metadata_variable_name`` introduced in the prior commit — the
delegation pulled ``RouteChecks`` into the same cycle, and the call
site reuses the resolved route for its other branches, so inlining
the substring check is both cycle-free and avoids a redundant second
``get_request_route`` call.
Comment in test_proxy_routes.py acknowledges that the two MCP table
entries exercise ``get_request_route`` directly rather than the full
production handler (which needs ASGI scope + MCP state to invoke).
---------
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: user <70670632+stuxf@users.noreply.github.com>
* fix(team): keep team_alias cache in sync on _cache_team_object writes (#28737)
* fix(team): keep team_alias cache in sync on _cache_team_object writes
_cache_team_object wrote only to the team_id:<id> cache key, but the
JWT auth path that uses team_alias_jwt_field reads from a separate
team_alias:<alias> key (get_team_object_by_alias caches under both
keys on miss, but reads only the alias-keyed one). After any
team-mutation endpoint (team_model_add, team_model_delete,
update_team, the two access-group writes) the team_id cache was
refreshed but the team_alias cache stayed stale until TTL — JWT
callers using team_alias_jwt_field kept seeing the pre-mutation
team for the full cache window.
Mirror the write under the alias key inside _cache_team_object so
every existing caller stays in sync without further changes. Skip
the alias write when team_alias is None/empty so we don't collide
across alias-less teams.
Surfaced testing the LIT-3244 cherry-pick on patch/1.86.0: the
LIT-3244 fix correctly invalidated the team_id cache but the
customer's JWT used team_alias_jwt_field, so they kept hitting the
stale alias-keyed entry.
* fix(team): delete (not overwrite) team_alias cache on _cache_team_object
The prior shape of this PR wrote both team_id:<id> AND team_alias:<alias>
from _cache_team_object. team_alias is NOT unique in the schema
(no @unique on LiteLLM_TeamTable.team_alias), and get_team_object_by_alias
enforces uniqueness on its own DB-fetch path (len(teams) > 1 raises).
Writing the alias-keyed cache from the generic refresh path bypassed
that check: a team admin renaming their team to collide with another
team's alias could silently overwrite the cached team for JWT-by-alias
auth, swapping the resolved team under that alias for the cache window.
Switch the alias-keyed operation from a write to a delete (mirroring
the dual-cache delete pattern in _delete_cache_key_object). After every
team write, the next JWT-by-alias reader cache-misses and falls through
to get_team_object_by_alias, which (a) re-fetches the fresh team from
DB, closing the LIT-3244 staleness gap that motivated this PR, and
(b) enforces alias uniqueness before populating either cache key.
team_id:<id> writes are unchanged — team_id is the table PK and is
guaranteed unique.
Surfaced in veria-ai review on #28739.
* fix(managed-files): anchor model_id regex so it doesn't match llm_output_file_model_id
extract_model_id_from_unified_id used `re.search(r"model_id,([^;]+)", ...)`
which substring-matches the `model_id,` inside the file-ID encoding's
`llm_output_file_model_id,<deployment_uuid>` field. parse_unified_id
then fed that deployment UUID back into the auth path as a model
candidate via _extract_models_from_managed_resource_id, and every
team-BYOK file attach 403'd with:
team not allowed to access model. This team can only access
models=['openai/*']. Tried to access <deployment-uuid>
The team's models list correctly contains the public name (`openai/*`)
that target_model_names matches, but the bogus UUID candidate fails
the wildcard check first.
Anchor the regex to a field boundary (`(?:^|;)model_id,`) so it
matches the legitimate top-level `model_id,<value>` field on
vector_store unified IDs and skips substring matches inside other
fields. File-IDs (which have no top-level `model_id` field) now
return None and contribute no spurious UUID candidate.
Surfaced reproducing LIT-3244 on patch/1.86.0 with the customer's
exact flow: team with openai/* BYOK deployment, JWT-scoped user,
POST /v1/vector_stores/{id}/files attaching a file uploaded with
target_model_names=openai/gpt-4o.
* fix(proxy): hydrate wildcard discovery credentials (#28284) (#28822)
* fix(proxy): hydrate wildcard discovery credentials
* fix(proxy): constrain wildcard credential hydration
Co-authored-by: Dibyo Mukherjee <dibyo@adobe.com>
* ci: add daily oss-agent-shin branch creation workflow (#28829)
Creates litellm_oss_agent_shin_MM_DD_YYYY from main every day at 00:00 UTC.
Lets us retarget oss-agent-shin fork PRs onto a canonical branch so CircleCI runs with secrets, without granting the agent write access.
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Ishaan Jaffer <ishaanjaffer0324@gmail.com>
* test(proxy): add harness for proxy_server.py behavior-pinning (#28827)
* test(proxy): add harness for proxy_server.py behavior-pinning
Creates tests/test_litellm/proxy/proxy_server/ with:
- conftest.py: 11 shared fixtures (app, client, mock_prisma, auth_as,
mock_router with parametrized response builders, normalize, etc.)
- _coverage_check.py: per-PR coverage gate (line + branch) against a
baseline, self-selects target by inspecting which placeholder files
have been filled
- _pin_check.py: AST-based gate that verifies every pin-list item has
>=1 happy + >=1 error test with a real assertion (no status-only)
- test_harness_smoke.py: 19 smoke tests covering every fixture +
both scripts end-to-end
- 26 placeholder test files (one docstring each) reserved for
follow-up PRs per the directory ownership in the Notion plan
- .coverage_baseline pinned at 0% so future PRs measure deltas
against new-tests-only and aren't entangled with the broader
scattered test suite
Adds a dedicated proxy-server job to test-unit-proxy-endpoints.yml
so this directory's runtime + coverage are tracked independently.
Plan: https://www.notion.so/36c43b8acdab81ee845fd5365128a2fc
* ci(proxy-endpoints): allow workflow_dispatch
Lets the workflow be triggered manually on a branch via
`gh workflow run`, which is needed for the verify-first
flow on workflow changes before opening a PR.
* test(proxy): address review feedback on proxy_server harness
- conftest.py: anchor sys.path insert to __file__ (Path(__file__).resolve().parents[4])
instead of CWD-relative os.path.abspath("../../../../") which resolved
to the wrong directory when pytest is launched from the repo root.
- _coverage_check.py: actually read .coverage_baseline and use it as
the floor (line_min = max(target, baseline)). Closes the gap between
the PR description's "delta semantics" and what the script was doing.
With baseline=0.0 today this is a no-op; future PRs that update the
baseline cause regressions (test deletions etc.) to trip the gate
even if the static PR target is still met.
- _pin_check.py: drop unreachable startswith("_") guard
(test_*.py glob never yields underscore-prefixed names) and read
each test file once instead of twice.
* feat(openai): apply regional-processing cost uplift for EU/US data residency (#28626)
* feat(openai): apply regional-processing cost uplift for EU/US data residency
OpenAI charges a 10% uplift on the latest GPT models when requests are
served from a regionalized hostname (eu./us.api.openai.com). Infer the
region from `api_base`, expose it on `kwargs["litellm_params"]["data_residency"]`,
and multiply the computed cost by a per-model
`regional_processing_uplift_multiplier_<region>` field.
https://claude.ai/code/session_012ebH44s7ohYxjoix5CXzTW
* test: allow regional_processing_uplift_multiplier_{eu,us} in model_prices schema
* fix(cost): tighten data_residency inference and restore model_cost in tests
- Only infer OpenAI data_residency when custom_llm_provider == "openai";
drop the implicit None fallback so non-OpenAI callers can't accidentally
pick up a regional tag from a stray OpenAI hostname.
- _local_model_cost_map fixture now snapshots and restores
litellm.model_cost and LITELLM_LOCAL_MODEL_COST_MAP so tests don't leak
state across the session.
* refactor(openai): move data_residency helper under llms/openai
* fix: thread data_residency through realtime stream cost calculation
Co-authored-by: Yassin Kortam <yassin@berri.ai>
* fix(cost): thread data_residency through batch_cost_calculator
Apply the OpenAI regional-processing uplift multiplier to retrieve_batch
cost paths so Batch API requests served via eu./us.api.openai.com are
priced at the same uplifted token rates as completions/transcriptions.
* refactor(openai): encapsulate provider check inside infer_openai_data_residency
Move the custom_llm_provider == "openai" guard from get_litellm_params
into the helper itself so the core utility no longer carries
provider-specific dispatch logic. Callers pass through the provider
unconditionally; the helper returns None for any non-OpenAI provider.
* fix(responses): thread data_residency through Responses logging params
The Responses API paths build their logging litellm_params dict after
provider resolution but did not include data_residency, so cost calc
saw None even when the effective api_base was a regional OpenAI host.
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Yassin Kortam <yassin@berri.ai>
---------
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: ryan-crabbe-berri <ryan@berri.ai>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: ryan-crabbe-berri <ryan-crabbe-berri@users.noreply.github.com>
Co-authored-by: milan-berri <milan@berri.ai>
Co-authored-by: Mateo Wang <277851410+mateo-berri@users.noreply.github.com>
Co-authored-by: Mateo <mateo@Mateos-MacBook-Pro.local>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Co-authored-by: Yassin Kortam <yassin@berri.ai>
Co-authored-by: Yassin Kortam <yassinkortam@Yassins-MacBook-Pro.local>
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: user <70670632+stuxf@users.noreply.github.com>
Co-authored-by: Krrish Dholakia <krrish+github@berri.ai>
Co-authored-by: Dibyo Mukherjee <dibyo@adobe.com>
Co-authored-by: ishaan-berri <155045088+ishaan-berri@users.noreply.github.com>
Co-authored-by: Ishaan Jaffer <ishaanjaffer0324@gmail.com>
* Revert "merge main (#28839)"
This reverts commit
|
||
|
Mateo Wang
|
b60d4677cd
|
fix(vertex_gemma): strip context_management from request body (#28438)
Vertex AI Gemma's chatCompletions wrapper does not understand the context_management parameter (an Anthropic / OpenAI Responses API concept). When callers route this field to a Gemma deployment (e.g. through allowed_openai_params or proxy passthrough), the upstream endpoint would reject the request with an unknown-field error. Drop context_management in VertexGemmaConfig.transform_request, matching the existing pattern used for stream and stream_options. Adds a direct transform_request unit test plus an acompletion-level test that exercises the realistic allowed_openai_params path. Co-authored-by: Cursor Agent <cursoragent@cursor.com> Co-authored-by: Mateo Wang <mateo-berri@users.noreply.github.com> |
||
|
Sameer Kankute
|
b7e978a5c3
|
Litellm oss staging 04 21 2026 2 (#26569)
* fix(bedrock): use model info lookup for output_config support instead of hardcoded check Replace hardcoded _is_claude_4_6_model() string matching with supports_output_config flag in model_prices_and_context_window.json, accessed via _supports_factory(). This follows the project's established pattern for model capability checks (per AGENTS.md rule #8). Bedrock Invoke now conditionally preserves output_config for models that declare supports_output_config=true (currently Claude 4.6 models), while stripping it for older models to avoid request rejection. Ref: https://github.com/BerriAI/litellm/issues/22797 * fix(vertex_ai): single-flight credential refresh to prevent thundering herd (#26024) * fix(vertex_ai): single-flight credential refresh to prevent thundering herd When GCP credentials expire under high concurrency, all requests simultaneously call credentials.refresh() via asyncify, saturating the 40-thread anyio pool and blocking the proxy for 20+ seconds. This adds: - Per-credential asyncio.Lock in get_access_token_async for single-flight refresh (1 coroutine refreshes, others wait on the lock) - Background refresh when token_state is STALE (usable but near expiry), returning the current token immediately with zero added latency - threading.Lock on the sync get_access_token path - Uses google-auth's TokenState enum (FRESH/STALE/INVALID) instead of reimplementing expiry logic Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: address PR review comments - Use asyncio.create_task() instead of deprecated get_event_loop().create_task() - Track in-flight background refresh tasks to prevent duplicate refreshes when multiple STALE-path callers pass through the lock before the first background task completes - Add token validation in the STALE branch (consistent with FRESH/INVALID) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: lazy-import TokenState to avoid breaking when google-auth is not installed Also extract helper methods to bring get_access_token_async under the PLR0915 statement limit (50). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: apply Black formatting to test file and update uv.lock Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: remove user-provided project_id from log messages (CodeQL log injection) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: avoid leaking token value in error message, log type instead Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: restore uv.lock to match litellm_oss_branch Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: remove project_id from remaining log message (CodeQL log injection) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: remove remaining project_id from log and error messages Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: reuse cached credentials in VertexAIPartnerModels (#26065) * fix: reuse cached credentials in VertexAIPartnerModels instead of creating new VertexLLM per request VertexAIPartnerModels.completion() was creating a throwaway VertexLLM() instance on every call to get an access token, bypassing the credential cache inherited from VertexBase. This caused a fresh token fetch for every single request, adding significant latency overhead. Fix: call super().__init__() to initialize VertexBase's credential cache, and use self._ensure_access_token() instead of a new VertexLLM instance. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix: apply same credential caching fix to VertexAIGemmaModels and VertexAIModelGardenModels Same bug as VertexAIPartnerModels: both classes had `pass` in __init__ instead of `super().__init__()`, and created throwaway VertexLLM() instances per request instead of using self._ensure_access_token(). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(fireworks): add glm-5p1 metadata and parallel_tool_calls (#26069) * fix(chatgpt): preserve responses routing and recover empty output (#25403) (#26219) - preserve existing shared backend `mode` when router deployment registration reuses a provider/model key already in `litellm.model_cost` (prevents alias with `mode: chat` from downgrading shared `chatgpt/gpt-5.4` from `responses` to `chat` and triggering 403s on /v1/chat/completions) - teach the ChatGPT Responses parser to recover `response.output_item.done` entries when `response.completed.output` is empty - add defensive /responses -> /chat/completions bridge fallback that reconstructs output items from raw SSE when `raw_response.output` is empty - regression coverage for shared alias routing, empty completed.output parsing, and SSE bridge recovery Closes #25403 Co-authored-by: afoninsky <andrey.afoninsky@gmail.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(deps): relax core runtime dependency pins from exact == to ranges When litellm migrated from Poetry to uv (PR #24905, v1.83.1), the core dependency specifications in pyproject.toml changed from Poetry bare-version strings (e.g. openai = "2.30.0") to PEP 621 exact pins (openai==2.24.0). Poetry bare-version strings are actually caret ranges (^X.Y.Z == >=X.Y.Z,<X+1), but PEP 621 == is exact. This means every downstream package that installs litellm as a library dependency is now forced to downgrade aiohttp, pydantic, openai, click, and 8 other common packages to exact old versions. Fix: restore range specifiers for the 12 core runtime dependencies. The optional extras (proxy, proxy-runtime, etc.) are consumed primarily by Docker images where exact pins are appropriate and are left unchanged. The uv.lock file continues to provide exact reproducibility for Docker builds and CI. Fixes: #26154 * Add Rubrik as officially-supported guardrail plugin (#25305) * Add Rubrik as officially-supported guardrail plugin Adds tool blocking and batch logging integration with an external Rubrik webhook service. The plugin validates LLM tool calls against a policy service (fail-open on errors) and batch-logs all requests/responses. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Update Rubrik docs: config.yaml as primary, env vars as fallback Restructures the Quick Start to present config.yaml as the recommended approach with tabbed UI, and environment variables as an alternative fallback. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Add Rubrik env vars to config_settings reference Fixes documentation validation by adding RUBRIK_API_KEY, RUBRIK_BATCH_SIZE, RUBRIK_SAMPLING_RATE, and RUBRIK_WEBHOOK_URL to the environment settings reference table. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Add fallback message when blocking service returns empty explanation Prevents whitespace-only violation message when the tool blocking service blocks tools but returns an empty content field. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * feat(ocr): add Reducto parse OCR support (#26068) * feat(ocr): add Reducto parse OCR support * fix(reducto): address OCR review feedback * chore: refresh uv lockfile * Revert "chore: refresh uv lockfile" This reverts commit |
||
|
Sameer Kankute
|
fecf212d70
|
fix(vertex_ai): omit function_call id on Vertex Gemini 3.5+ tool turns (#28324)
* fix(vertex_ai): omit function_call id on Vertex Gemini 3.5+ tool turns Vertex AI rejects `id` on function_call/function_response parts; only Google AI Studio accepts it for Gemini 3.5+ strict tool matching. Co-authored-by: Cursor <cursoragent@cursor.com> * Update litellm/llms/vertex_ai/gemini/vertex_and_google_ai_studio_gemini.py Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * fix(vertex_ai): forward custom_llm_provider in context caching Pass custom_llm_provider through to _gemini_convert_messages_with_history in the context caching path so Gemini 3.5+ tool-call `id` forwarding behaves consistently between cached and non-cached completions on Google AI Studio. Co-authored-by: Claude <claude@anthropic.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: Claude <claude@anthropic.com> |