* chore(ui): remove Agent Platform announcement bell from navbar
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* feat(ui): highlight Auto Router in the navbar announcement
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(ui): open Auto Router docs link in a new tab
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* Revert "fix(ui): open Auto Router docs link in a new tab"
This reverts commit 3be861bafe.
* chore(ui): title the navbar announcement LiteLLM Auto Router
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: Mubashir Osmani <mubashir@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Prettier flagged the two test files added while fixing review findings.
Migrating these components off Ant Design also retired the lint suppressions
they carried, so prune those 15 entries and leave the unrelated ones for the
PRs that made them stale.
antd's Collapse kept the panel mounted once opened, so a tool a user had
expanded stayed expanded after closing and reopening Tools. Base UI renders
only the open branch, so the migration silently reset every ToolItem.
The regression test passes against the antd original, fails against the
migration without keepMounted, and passes with it.
* fix(proxy/batches): stop forwarding custom_llm_provider twice in list and cancel
The model-routing branches of list_batches and cancel_batch passed
custom_llm_provider as an explicit kwarg while also leaving it inside the dict
they splat, so every such call raised "got multiple values for keyword argument
'custom_llm_provider'" and returned a 500.
list_batches SCENARIO 2 called data.update(credentials) but never removed
custom_llm_provider before litellm.alist_batches(custom_llm_provider=..., **data);
it now uses prepare_data_with_credentials, the same helper the create and
retrieve branches already use, which pops it out.
cancel_batch SCENARIO 3 resolved the provider with
`provider or data.pop("custom_llm_provider", None) or ...`, so when the path
param provider was set the pop short-circuited and a body custom_llm_provider
stayed in data and collided with the explicit kwarg. The body value is now
popped unconditionally before the fallback chain, so the path param wins cleanly
and data no longer carries a duplicate.
Both paths already had strict-xfail regression tests documented "remove when
fixed"; those markers are dropped so the tests now guard the fix.
Signed-off-by: Anas Khan <83116240+anxkhn@users.noreply.github.com>
* fix(proxy/files): avoid duplicate custom_llm_provider in list
Signed-off-by: Anas Khan <83116240+anxkhn@users.noreply.github.com>
---------
Signed-off-by: Anas Khan <83116240+anxkhn@users.noreply.github.com>
InputCard and OutputCard located SectionHeader's copy button by querying for a
descendant with aria-label="copy", which is the antd CopyOutlined icon. That
selector reaches into SectionHeader's internals, so migrating it off antd left
copyButton undefined and failed four tests.
getByRole("button", { name: /copy/i }) is green against both the antd and the
shadcn SectionHeader, verified by running these two files against each.
The key source trigger rendered the stored value, so the playground showed
session and custom instead of Current UI Session and Virtual Key. Name the
selected option on the trigger.
Clearing the key while models were loading left the selector disabled for
good: the in-flight load skips its reset once cancelled, and the branch that
handles an empty key returned without clearing the loading flag, so nothing
put it back. Clear it on that path too.
* fix(ui): restore playground model filtering by endpoint
Bring back the prior Chat model dropdown filter (including chat models
on responses/anthropic/interactions and image models on image_edits), and
map mode realtime so the realtime endpoint only lists compatible models
* fix(ui): exclude unknown model modes from playground endpoint filters
Modes outside ModelMode (batch, rerank, ocr, etc.) must not collapse to
chat-compatible, or conversational endpoints surface unusable models
* feat(ui): add shared vercel-style playground chat composer (#36131)
* feat(ui): adopt vercel-style chat composer for playground
Replace the compact single-line input with a PromptInput-style composer:
taller auto-growing textarea, rounded card shell, footer tools, and
stop button while a request is in flight
* style(ui): strengthen playground chat composer border and shadow
Make the shared chat input stand out with a fuller border, layered
shadow, and a slightly stronger focus ring
* fix(ui): size chat composer textarea with CSS field-sizing
Drop direct el.style.height mutation in favor of field-sizing:content
* fix(ui): keep the chat composer out of a nested form and focus its textarea
The composer wrapped everything in a native form, so MCP mode nested Ant
Design's tool-arguments form inside it, which is invalid HTML and let Enter
hit either form. The footer also relied on InputGroupAddon focusing the first
input in the group, which is the hidden file input from the attach controls
rather than the message textarea.
Drop the outer form and submit from the send button directly, and have the
addon focus the element marked as the group's control.
* refactor(ui): reuse the endpoint compatibility check when a model is picked
The endpoint guard added upstream duplicated the compatibility families this
PR introduces, so point it at isModelCompatibleWithEndpoint instead. Filtering
also means an incompatible model is no longer offered for an endpoint, so the
test that picked one now asserts it is absent.
* fix(ui): match the image-edit model mode the backend actually sends
model_prices_and_context_window.json labels these models image_edit, but the
mode enum spelled it image_edits, so once unknown modes started being filtered
out every image-edit model vanished from the playground, /v1/images/edits
included. The endpoint key keeps its own spelling.
The compatibility tests stubbed getEndpointType with a hand-written map that
repeated the same wrong spelling, which is how this stayed hidden, so they now
run against the real mapping.
Picking a model reset the endpoint from its mode unconditionally, so choosing
a chat model while on /v1/responses, /v1/messages or interactions bounced the
playground to /v1/chat/completions. Only switch when the current endpoint
cannot serve the picked model.
The temperature and max-token boxes parsed and clamped on every keystroke, so
a decimal lost its point and clearing the field snapped to a bound. They are
now text fields with a numeric input mode that hold what was typed and clamp
on blur; the sliders beside them still give the stepped control.
The image-edit and transcription areas invited a drag but had no drop
handlers after the Ant Design Dragger came out, so drops did nothing. Wire
drop through the same validation the file picker uses.
* feat(complexity_router): calibrate the classifier rubric with worked examples
The built-in rubric stated its tier boundaries as prose alone, and prose
calibrated to consumer chat puts "non-trivial code, multi-step technical work"
at the top of the scale. That is the median request in developer and agent
traffic, so ordinary engineering read as top-tier and the router paid for the
most expensive model on it.
Adds calibration examples to the rubric, selected by a new
classifier_llm_config.rubric preset. The agentic preset (now the default)
anchors routine installs, builds, multi-file edits, and standard debugging at
MEDIUM; the chat preset omits those anchors for deployments serving only
conversational traffic. Both share the same tier criteria, the trust-boundary
paragraph, and the context-window closing line, so this moves where the
boundary sits without changing the taxonomy.
Both presets render byte-identical to the strings a prompt sweep scored, and a
test pins that, so the measured accuracy describes what a router sends.
* feat(ui): pick the classifier rubric preset on an auto-router
Adds a Rubric dropdown to the auto-router's classification panel, so the
agentic and chat presets are selectable rather than config-file only. The
prompt editor prefills from the selected preset, since prefilling agentic text
for a router on chat would show examples its classifier never receives.
The picker is disabled while a custom prompt is set, and the payload builder
drops the preset in that case: a custom prompt is the classifier's whole system
role, so the backend rejects the two together. The builder records the default
preset explicitly, so a later change to which preset is default cannot silently
move an existing router.
* fix(complexity_router): mark an unchosen rubric preset with None, not model_fields_set
The mutual-exclusion check read model_fields_set to tell an explicit preset
from the default. That flag does not survive serialization, and this config is
dumped and handed straight back to ComplexityRouter by /auto_router/test_routing,
where a dump re-states every field. So a custom-prompt classifier saved fine and
then failed validation on preview, rejecting on the second pass what it accepted
on the first.
The preset is now optional, with None meaning the default, matching how None
already means the built-in rubric for system_prompt on the same model. The
default lives in one place, DEFAULT_RUBRIC_PRESET, resolved where the prompt is
assembled. The dashboard stops sending a copy of the default it displays, so a
router nobody configured follows the default rather than pinning today's value,
and UI-built routers behave the same as hand-written config.
Regenerates schema.d.ts, which was left stale by an earlier description edit.
* feat(complexity_router): grandfather existing routers onto the uncalibrated rubric
An unset preset now means LEGACY, the rubric exactly as it shipped before
calibration examples existed, so upgrading cannot move the tier decisions or the
bill of a router that is already running. Config-file routers get this for free
since they name no preset, and a stored config that never had one reads the same
way.
New routers still get the calibrated rubric: switching a classifier to LLM
stamps the agentic preset, because a classifier being configured for the first
time has no prior tier behaviour to preserve. The picker offers legacy so an
existing router's state is representable and opening the form cannot silently
upgrade it.
Each preset is pinned byte-identical to the text the prompt sweep scored,
legacy included, which is what proves an existing router's prompt did not move.
Also collapses the preset data from a NamedTuple with group wrappers and
per-preset frozensets into plain text blocks in a MappingProxyType, matching how
the tier criteria next to it are already stored: 21 lines of prompt text no
longer cost 190 lines of constructors. Tiers are format placeholders so
tier_labels still reach the examples.
* refactor(complexity_router): name the field classification_rubric
`rubric` alone did not say what it selects, and the field sits beside
`system_prompt`, which genuinely is the whole classification prompt. The name
now says which of the two an operator is reaching for: the rubric the built-in
prompt is assembled from, not the prompt itself.
Renames the config field, the query param, the enum, and the dashboard label to
match, and moves the preset text to classification_rubrics.py.
* test(ui): set the preset the mutual-exclusion case is meant to drop
The rename left classification_classification_rubric in the custom-prompt case,
so its input never carried a preset and the assertion held for the wrong reason:
it proved an absent preset stays absent, not that a set one is dropped. A
normalizer that forwards the preset whenever one is set passed with the typo and
fails without it.
tsc reports the typo as TS2353; the earlier sweep grepped for the source file
and not the test, so it went unseen.
* test(ui): scope the role-gate assertions to each page's own endpoint
The memory, workflows, and guardrails-monitor page tests asserted that a denied
role fires no request at all. Their names, and the assertion on the very next
line, say the intent is narrower: the page must not fetch its own data.
Resolving whether a caller is an org admin goes through /organization/list for
every role, since deciding org-admin-for-any-org needs the list, and the route
scopes rows per caller. That legitimate request fails a blanket no-fetch
assertion, so all three files went red on staging for a reason unrelated to
what they test.
Drops the blanket assertion and keeps the scoped one. Bypassing the gate in
memory/page.tsx still fails five tests, so the narrower assertion continues to
catch a genuinely broken gate.
* fix(complexity_router): document that an unset rubric keeps the legacy prompt
The field said 'Leave unset for agentic' while an omitted rubric resolves to
LEGACY, so the OpenAPI schema an operator reads promised calibrated routing
where they got the uncalibrated one.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
System prompts (harnesses, tools, framework boilerplate) are session-wide
constants identical across all requests. Scoring them saturates keyword-match
signals and produces false-positive high-complexity classifications on
trivial utterances like 'hi', routing them to expensive models (sonnet/opus)
instead of tier-1 haiku. A real ~1.6KB CLI-agent harness alone supplied
5 codePresence + 2 technicalTerms matches, overshadowing user signal.
Rescope four scoring dimensions (codePresence, technicalTerms, simpleIndicators,
multiStepPatterns) from full_text (system + user) to user_text (user only).
reasoningMarkers was already scoped this way. This returns 0.63 of the weight
budget to text that actually varies per-request.
Now that every dimension scores user_text only, _score_keyword_match's
disclosable_text param is redundant -- it existed solely to let the signal
name terms matched in the caller's own message while withholding terms
matched only in the (invisible-to-the-caller) system prompt. With no more
system-prompt text in scope, text and disclosable_text were identical at
every call site, so the param is dropped and the function collapses to a
single text argument.
Add mutation-proven regression test: trivial 'hi' message with realistic
Claude Code agent system prompt now routes to haiku tier-1 (not sonnet).
- Unfixed: haiku -> sonnet (bug)
- Fixed: haiku -> haiku (correct)
Invert three pre-existing assertions in TestSignalsNeverQuoteTheSystemPrompt
to capture the corrected behavior: system-prompt-only terms produce no signal.
Co-authored-by: Claude <noreply@anthropic.com>
* test(ui): decouple usage table test from antd
* refactor(ui): migrate usage tables to shared DataTable
* test(ui): preserve data utility exports in usage tests
The summary model access and budget checks read team_id, user_id,
project_id, and end_user_id via getattr with a None default so duck-typed
auth objects without those attributes keep working
Types 28 files with Protocols, TypedDicts, and Pydantic validation in place
of Any, cutting basedpyright reportAny by 974 and reportExplicitAny by 261
(1411 errors total across 48 rules), and ratchets the basedpyright, ruff
strict, and type discipline budgets down to match
langfuse_* request headers land in metadata as strings, but the trace path reads
mask_input/mask_output with a bare truthiness check and iterates update_trace_keys
directly. A header saying mask_input: false redacted the payload it was asked to
keep, and update_trace_keys was walked one character at a time so every requested
key silently failed to match
The template rendered affinity and tolerations but never nodeSelector, so a
values file that pinned the chart to a node pool got the gateway and every
subchart placed correctly while the migration Job silently fell through to
whatever the cluster's default pool was.
That is worse than an outright failure. On EKS Auto Mode the default pool hands
out 3 GiB nodes and the migration container needs roughly 3.6 GB, so the Job
was OOM-killed on a pool it was never meant to run on, while the values file
that would have placed it on a large enough node looked correct.
The new test fails against the old template with "unknown path
spec.template.spec.nodeSelector".
* fix(guardrails): scan and re-emit raw Anthropic SSE streams in the bedrock post-call hook
* fix(guardrails): keep upstream id and model on a blocked Anthropic stream
* fix(guardrails): deliver a blocked Anthropic stream as an error frame
* fix(guardrails): deliver an unscannable Anthropic stream as an error frame
* fix(guardrails): emit the guardrail block detail as JSON in the stream error frame
* fix(guardrails): deliver an Anthropic block through the shared block-SSE builder
* fix(guardrails): keep the shared SSE assembler behavior-identical for existing callers
* fix(guardrails): keep the stream error message a string and drop an unreachable branch
* chore(guardrails): drop a comment that repeated its own docstring
* fix(guardrails): let bedrock service failures keep their status instead of framing them as blocks
* fix(guardrails): key the streamed block decision on status, not detail shape
InvokeGuardrailChecks details a Mapping on its 500 for an unparseable response,
so a detail-shape test read that outage as a policy block and framed it as a 200
guardrail_error. Both block sites raise 400, so gate on the status too.
* refactor(guardrails): narrow the SSE error-frame helper to the input it actually takes
Both callers pass a string, so the Mapping overload and its json.dumps branch
were unreachable. Folds the block branch's narrative comment into the rebind
suppressions that already carry a reason.
