scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-17 23:51:30 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 67
49820 commits 2515 branches 1410 tags 3 GiB
Commit graph

49820 commits

This branch
This branch
All branches
Author SHA1 Message Date
kerry
10aef224e7 style(anthropic): apply repository conventions to the missing-usage change 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:34:28 +00:00
kerry
91c8d1cdc1 chore: merge origin/main into litellm-providers/price-sync 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:30:24 +00:00
kerry
d1aa37b0a8 merge main into litellm_fix_fireworks_cost_components 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:28:29 +00:00
Yuneng Jiang
45d5e6b833
fix(e2e): start cache CI service and count bypass calls 2026-09-15 18:26:44 -07:00
kerry-berri
171888b716
Merge pull request #41298 from BerriAI/litellm_drop_remaining_vendor_fact_pins 
test: drop remaining tests that pin cost-map vendor facts
 2026-09-15 18:25:14 -07:00
kerry
a6d2332f7a test(responses): drive the usage-estimate failure path without patching litellm 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:24:35 +00:00
kerry
5a9fe56aff fix(responses): count custom-tool and MCP argument deltas in the streamed usage fallback 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:24:05 +00:00
yucheng-berri
a9a654cb60
Merge pull request #38241 from BerriAI/litellm_agent365_mcp_guardrail 
feat(guardrails): add Microsoft Agent 365 MCP tool-call guardrail
 2026-09-15 18:21:46 -07:00
Yuneng Jiang
f214555068
fix(e2e): expect models filters to persist after reload 2026-09-15 18:21:43 -07:00
kerry
44a6d19889 test(gemini): drop review narration from the wrapper test docstring 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:14:37 +00:00
Yuneng Jiang
4587dbeae9
feat(e2e): cache exact provider responses for 24 hours 2026-09-15 18:14:18 -07:00
kerry
1484fd7600 fix(responses): keep the usage estimate best-effort when token counting raises 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:13:26 +00:00
yuneng-jiang
113a416619
Merge pull request #41319 from BerriAI/litellm_fix_ui_onboarding 
fix(e2e): onboard dashboard users through invitations
 2026-09-15 18:06:52 -07:00
yucheng
f66ffc387e fix(guardrails): only honor the judge call-origin stamp on logging_only in llm_as_a_judge 
On pre_call, during_call and post_call the hook data is the client request body, so a
client-supplied litellm_params.metadata.internal_call_origin must not skip enforcement.
Type the during_call helper's request payload as dict[str, object]

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
kerry
801f6a92b4 test(gemini): assert the served modelVersion reaches the assembled stream through CustomStreamWrapper 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
dd4829398a fix(guardrails): log the configured mode when logging_only is mixed with an enforcing llm_as_a_judge mode 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
10506ab904 fix(guardrails): judge the whole latest user turn, run every during_call guardrail, log combined modes as configured 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
1b8b17cc8a fix(guardrails): judge only the latest request turn in pre_call and during_call llm_as_a_judge 
The request-side prompt told the judge to focus on the most recent user turn but the text under review was every extracted request message joined together, so a multi-turn request with an off-topic earlier turn and an on-topic latest turn scored 50 and was blocked. Request-side judging now evaluates the last extracted request text (after the configured message scoping) and passes the full role-labelled conversation only as context. Response-side judging still evaluates all extracted response text

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
e9357d9a6f fix(guardrails): stop logging_only llm_as_a_judge from judging its own judge calls 
Judge sub-calls now carry the internal_call_origin metadata stamp and the
guardrail skips any logged call bearing it, so a logging_only judge no longer
recurses into an unbounded chain of judge requests

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
a658f20005 test(guardrails): grant premium_user for the tagged Mode case in llm_as_a_judge mode-shape test 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
4d1330ea0b test(guardrails): drop callback-manager patch from llm_as_a_judge mode-shape test 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
28bd00a004 fix(guardrails): label llm_as_a_judge logging_only verdicts with their mode 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
b93f22265c fix(guardrails): give the request-side judge role-labelled conversation context 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:53 +00:00
yucheng
6fccf00d93 fix(guardrails): keep list and tagged mode shapes for llm_as_a_judge 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:52 +00:00
yucheng
0b34800501 fix(guardrails): resolve llm_as_a_judge request-side log mode from event_hook lists, drop unused alias 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:52 +00:00
yucheng
77a6327675 feat(guardrails): support pre_call and during_call modes for llm_as_a_judge 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:05:52 +00:00
kerry
de9aa48cd6 fix(responses): count multimodal input and tool-call output in the streamed usage fallback 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 01:02:29 +00:00
Tin Chi Lo
5aa2adbacd fix(proxy): preserve Anthropic pricing modifiers in router savings 2026-09-15 18:02:05 -07:00
mateo-berri
3c15f64fd4 fix(bedrock): make prompt caching work on the Nova InvokeModel route 
Nova InvokeModel rejects the standalone cachePoint blocks the shared Converse transform emits, so each one is folded into the block it caches and tool_config injection points are dropped before the transform runs, since this route has no tool caching to credit. Usage reads Bedrock's Count-suffixed cache keys and adds cached tokens into prompt_tokens, streaming routes every wrapped InvokeModel event through the Converse chunk parser and tolerates the missing totalTokens, and the Nova 1 cost-map entries gain cache_read_input_token_cost at a quarter of the input rate
 2026-09-15 18:01:21 -07:00
kerry
405a127838 fix(fireworks-ai): drop banned typing.cast to a suppressed import for the copied pricing entry 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:57:05 +00:00
mateo-berri
3a8679d3d8 fix(proxy): never forward the LiteLLM virtual key to Anthropic on the /anthropic passthrough 
The /anthropic/{endpoint} route forwarded every incoming header upstream, so
the header carrying the caller's LiteLLM virtual key (Authorization, x-api-key,
x-litellm-api-key, or the operator-configured key header) reached Anthropic and
was rejected there as an invalid credential, with or without a proxy-side
Anthropic key layered on top.

Share the Vertex credential-less header filter: drop the proxy-only credential
headers by name, drop the value that authenticated the caller (virtual key,
master key, or JWT) from Authorization / x-api-key, keep a caller's own
Anthropic credential, layer the proxy's Anthropic credential on top, and fail
with a clean 401 when neither the proxy nor the caller supplied one.

Resolves LIT-3550
 2026-09-15 17:54:58 -07:00
mateo-berri
b4f9e319fc fix(proxy): surface the upstream status code when a RAG query fails 2026-09-15 17:53:57 -07:00
yucheng
9bd3f7b885 Merge remote-tracking branch 'origin/main' into litellm_agent365_mcp_guardrail 2026-09-16 00:52:27 +00:00
mateo-berri
ba971d69d6 test(otel): type the providerless-call helpers and inject the logger provider 2026-09-15 17:51:28 -07:00
yucheng-berri
0b3e56448f
Merge pull request #41126 from BerriAI/litellm_custom_code_guardrail_identity 
fix(guardrails): resolve caller identity from metadata buckets in custom code guardrail
 2026-09-15 17:47:17 -07:00
yassin
14d239fc84 test(proxy): pin JWT scope denial message shape through auth exception conversion 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:46:59 +00:00
mateo-berri
8b0d28c219 Merge remote-tracking branch 'origin/main' into litellm_otel_gen_ai_system_none 
# Conflicts:
#	litellm/integrations/opentelemetry.py
#	litellm/integrations/opentelemetry_utils/gen_ai_semconv.py
#	tests/test_litellm/integrations/test_opentelemetry.py
 2026-09-15 17:46:41 -07:00
Yuneng Jiang
388eef4fbb
fix(e2e): onboard dashboard fixtures through invitations 2026-09-15 17:44:18 -07:00
kerry
5eab1feb20 fix(fireworks-ai): bill cache-write, reasoning, and audio tokens via the shared cost calculator 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:42:13 +00:00
Devin AI
232233f654 refactor(fireworks_ai): extract reasoning_effort mapping into helper 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:39:11 +00:00
Devin AI
a7a61db78d fix(fireworks_ai): flatten dict-form reasoning_effort to its effort string 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:37:55 +00:00
tin-berri
878716f806
Merge pull request #41326 from BerriAI/litellm_forecast_classifier_entitlement 
feat(router): limit unlicensed Capability and Fuse v2 routers to one each
 2026-09-15 17:36:22 -07:00
kerry
d5a36eb2ca fix(gemini): propagate provider modelVersion onto model responses 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:36:07 +00:00
kerry
ba6bcd747f chore: merge origin/main into test cleanup 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:33:25 +00:00
kerry
f5c1c82f81 fix(responses): estimate usage from text when streamed completed event omits usage 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:32:24 +00:00
Mateo Wang
aa710dc6a9
Merge pull request #35388 from BerriAI/litellm_session_total_duration 
fix(spend): sum multi-round session duration in logs UI
 2026-09-15 17:31:41 -07:00
Joshua Valluru
61e3b5ddae fix(mcp): persist OAuth credentials for rowless JWT admins 2026-09-15 17:31:07 -07:00
kerry-berri
daa98cd6ff
Merge pull request #41320 from BerriAI/litellm_gemini_fallback_generalizations 
feat(model_info): add provider-neutral Gemini 2.5+ chat baseline fallback generalization
 2026-09-15 17:28:58 -07:00
Yassin Kortam
7eeba69016
Merge pull request #41316 from BerriAI/litellm_nvidia_nim_infer_passthrough 
feat(proxy): add /nvidia_nim passthrough route for NIM object detection and OCR /v1/infer
 2026-09-15 17:28:25 -07:00
kerry
4657d43fb9 fix(anthropic): tolerate message_delta chunks without usage in streams 
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-09-16 00:27:09 +00:00