Skip to content

fix(cache): log hashed cache keys#29890

Merged
Sameerlite merged 1 commit into
BerriAI:litellm_oss_staging_080626from
fengjikui:litellm_fix_cache_key_hash
Jun 8, 2026
Merged

fix(cache): log hashed cache keys#29890
Sameerlite merged 1 commit into
BerriAI:litellm_oss_staging_080626from
fengjikui:litellm_fix_cache_key_hash

Conversation

@fengjikui

Copy link
Copy Markdown
Contributor

Relevant issues

Fixes #29414

Linear ticket

N/A

Pre-Submission checklist

  • I have added meaningful tests
  • My PR passes all unit tests on make test-unit
  • My PR's scope is as isolated as possible; it only solves 1 specific problem
  • I have requested a Greptile review by commenting @greptileai and received a Confidence Score of at least 4/5 before requesting a maintainer review

Delays in PR merge?

N/A

CI (LiteLLM team)

  • Branch creation CI run
    Link:

  • CI run for the last commit
    Link:

  • Merge / cherry-pick CI run
    Links:

Screenshots / Proof of Fix

I reproduced the reported debug-log shape by calling Cache.get_cache_key() with a long system prompt, tools, response_format, and messages. The returned cache key was already a 64-character SHA-256 key, but the Created cache key debug log printed the unhashed source material before this change

After this change, the same script shows the log line contains the final hashed key and the source material length, while the prompt marker is absent from logs

cache_key_len= 64
prompt_marker_in_logs= False
Created cache key: af820855c8cf25649dea1d7decef5da2401d6019181a39ae3a881f78bde3d1ef (source material length: 2559)

Validation run locally:

uv run python -m pytest tests/test_litellm/test_model_param_helper.py tests/test_litellm/caching/test_caching.py -q
5 passed in 0.10s

uv run ruff check litellm/caching/caching.py tests/test_litellm/caching/test_caching.py
All checks passed!

uv run ruff format --check tests/test_litellm/caching/test_caching.py
1 file already formatted

git diff --check

I did not run make test-unit locally because this is a focused cache logging change and the targeted cache-key regression passed

Type

Bug Fix
Test

Changes

Cache.get_cache_key() now emits the final hashed cache key in the Created cache key debug log instead of the raw source material used to derive the hash. The log keeps source material length so operators can still spot unexpectedly large cache-key inputs without writing prompt content, tool schemas, or response schemas into debug logs

A regression test covers long prompt material with tools and response_format. It asserts the returned key remains a SHA-256 value and the Created cache key debug log does not contain prompt material

@fengjikui

Copy link
Copy Markdown
Contributor Author

@greptileai

@greptile-apps

greptile-apps Bot commented Jun 7, 2026

Copy link
Copy Markdown
Contributor

Greptile Summary

This PR fixes a debug-log privacy issue in Cache.get_cache_key() where the Created cache key log line was emitting raw source material (messages, tool schemas, response formats) instead of the final SHA-256 hash. The log is now moved to after the hash and namespace steps, logging the final key and the source-material byte length.

  • litellm/caching/caching.py: Relocated the verbose_logger.debug call three lines later so it fires after _get_hashed_cache_key and _add_namespace_to_cache_key, replacing the raw string with the hashed key and adding a source material length field.
  • tests/test_litellm/caching/test_caching.py: New unit test that exercises get_cache_key with a long prompt marker, tools, and response_format, then asserts the returned key is a valid 64-char hex digest and that the marker string is absent from all Created cache key log records.

Confidence Score: 5/5

Safe to merge — the change is a three-line log relocation with no effect on cache key computation or lookup behavior.

The only touched logic is the placement of a debug log statement. The hash computation, namespace addition, and return value are all unchanged. The new test directly validates the fix and makes no network calls.

No files require special attention.

Important Files Changed

Filename Overview
litellm/caching/caching.py Moved debug log for "Created cache key" to after hash computation, now logs the final hashed key and source material length instead of raw source material
tests/test_litellm/caching/test_caching.py New regression test verifying the cache key debug log does not contain prompt/tool/response_format content, and that the returned key is a valid SHA-256 hex string

Reviews (1): Last reviewed commit: "fix(cache): log hashed cache keys" | Re-trigger Greptile

@greptile-apps

greptile-apps Bot commented Jun 7, 2026

Copy link
Copy Markdown
Contributor

Greptile Summary

This PR fixes a debug-logging privacy leak in Cache.get_cache_key(), where the Created cache key log line emitted the full raw source material (messages, tools, response_format) rather than the final SHA-256 hash. The fix reorders two lines so the log fires after hashing and after the namespace is applied, keeping source material length for diagnostics.

  • litellm/caching/caching.py: The single verbose_logger.debug call is moved two lines down (after _get_hashed_cache_key and _add_namespace_to_cache_key) and updated to log hashed_cache_key and len(cache_key) instead of the raw cache_key string.
  • tests/test_litellm/caching/test_caching.py: A new pure-local test asserts the returned key is a valid 64-char hex digest, the log contains that digest, and the log does not contain the injected prompt_marker.

Confidence Score: 5/5

The change is minimal and additive — it only reorders two existing lines and adjusts a single log call. Cache key computation and hashing are untouched; only the log output changes.

Both files are straightforward: a two-line reorder in the production code and a new local-only pytest that directly verifies the fixed behavior. There are no functional changes to the cache key derivation, hashing, or namespace logic. The test covers the exact regression described in the issue.

No files require special attention.

Important Files Changed

Filename Overview
litellm/caching/caching.py Moves the Created cache key debug log to after hashing so the logged value is the final SHA-256 key; adds len(cache_key) for operator diagnostics without leaking raw content.
tests/test_litellm/caching/test_caching.py New test file that validates the returned cache key is a 64-char SHA-256 hex digest and that the Created cache key debug log does not contain raw prompt material.

Reviews (2): Last reviewed commit: "fix(cache): log hashed cache keys" | Re-trigger Greptile

@codecov

codecov Bot commented Jun 7, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@Sameerlite Sameerlite changed the base branch from litellm_internal_staging to litellm_oss_staging_080626 June 8, 2026 12:16
@Sameerlite Sameerlite merged commit 189ce26 into BerriAI:litellm_oss_staging_080626 Jun 8, 2026
70 checks passed
mateo-berri pushed a commit that referenced this pull request Jun 8, 2026
* feat(bedrock_mantle): add SigV4/IAM auth to Responses API route (fixes #29665) (#29788)

* feat(responses): add default no-op sign_request to BaseResponsesAPIConfig

* feat(responses): call sign_request after body is final, send signed bytes when signed

* feat(bedrock_mantle): add SigV4 sign_request via composed BaseAWSLLM (bearer path)

* test(bedrock_mantle): cover SigV4 access-key, AssumeRole, body bytes, region/auth consistency

* feat(bedrock_mantle): defer auth to sign_request; validate_environment no longer requires bearer

* docs(bedrock_mantle): document SigV4 + Bearer auth on Responses route

* test(responses): cover fake-stream signing order and mantle bearer arg/env precedence

* fix(bedrock_mantle): wrap all botocore credential errors with both-paths guidance

* fix(bedrock_mantle): catch specific credential errors, not all BotoCoreError, so STS transport failures are not masked

* fix(bedrock_mantle): sign the compact Responses route too, not just create

* fix(github-copilot): route per-model on /v1/responses based on model info (#29747)

* feat(focus): add GCS destination for FOCUS export (#29751)

* test: add failing tests for FocusGCSDestination

* feat: add FocusGCSDestination reusing GCSBucketBase auth

* feat: register FocusGCSDestination in factory; export from __init__

* fix(focus): preserve GCS_PATH_SERVICE_ACCOUNT when service_account_json not in config

* style: apply Black formatting to gcs_destination and tests

* style: apply Black formatting to factory.py

* fix(bedrock): omit empty additionalModelRequestFields and system from Converse API payload (#29565)

Amazon Nova Pro (and other strict Bedrock models) return 400 Malformed input
request when additionalModelRequestFields: {} or system: [] are present in the
payload. Both fields are optional in CommonRequestObject (total=False) and must
be omitted rather than sent as empty structures.

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible in pass-through cost tracking (#29730)

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible

Azure OpenAI resources created via the newer "Azure AI Foundry" /
Cognitive Services pathway live on `*.cognitiveservices.azure.com`
subdomains, not the older `openai.azure.com`. Both are valid Azure
OpenAI surfaces in production today.

The OpenAI pass-through cost-tracking handler hard-codes only the older
hostname in five places (four `is_openai_*_route` methods on
OpenAIPassthroughLoggingHandler, plus is_openai_route on
PassThroughEndpointLogging). As a result, calls from newer Azure
deployments are silently classified as "not an OpenAI route", the
dispatch into the cost-tracking handler is skipped, and tokens/cost
never get extracted into LiteLLM_SpendLogs — the row gets written with
prompt_tokens=0, completion_tokens=0, spend=0, model='unknown'.

Reproduced 2026-06-04 against a real Azure OpenAI deployment on
`*.cognitiveservices.azure.com` proxied through LiteLLM v1.88.0.

Fix: factor the hostname check into a single helper
`_is_openai_compatible_host` listing all three recognized surfaces
(api.openai.com, openai.azure.com, cognitiveservices.azure.com), and
have all five call sites delegate to it. Purely additive — never
weakens recognition for the originally-supported hostnames.

Adds a test
`test_is_openai_route_recognizes_cognitiveservices_azure_com` that
exercises all four `is_openai_*_route` static methods against
`*.cognitiveservices.azure.com` URLs (positive cases per route + a
small cross-route negative to confirm route-specific path matching
still works on the new hostname).

Out of scope for this PR (separate followup):
  - `openai_passthrough_handler` calls chat/completions
    `transform_response` on Responses API payloads (`output:` not
    `choices:`), which throws inside the dispatch and drops the
    SpendLogs row entirely. Recognized + tracked separately.

* ci: trigger fresh run

Empty commit to re-run checks. The previous auth-and-jwt failure was
a transient HuggingFace Hub 429 rate-limit hitting tokenizer downloads
in tests/proxy_unit_tests/test_custom_tokenizer_bug.py — unrelated to
this PR's scope (hostname recognition in pass-through cost tracking).
No code change.

---------

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* fix(responses): preserve forced-function tool_choice name in Responses to Chat transform (#29812)

The Responses API forces a specific function with a top-level name
({"type": "function", "name": "X"}), but _transform_tool_choice only handled the
nested Chat Completions shape and fell through to returning "required" for the flat
form, silently dropping the function name and degrading a forced function call to
force-any-tool. Map the flat Responses shape to the nested Chat shape, keeping the
"required" fallback when no name is present.

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic (#29584)

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic

PR #20951 strips system blocks beginning with "x-anthropic-billing-header:" for
every Anthropic target. That block is how the first-party Anthropic API recognizes
Claude Code subscription (OAuth) traffic, so dropping it makes requests that carry
only that block, such as the auto-mode tool-safety classifier, fail with a
misleading 429 rate_limit_error; normal turns still work because they also carry
the "You are Claude Code" identity block.

Gate the strip behind should_strip_billing_metadata(), defaulting to False on the
first-party AnthropicConfig and AnthropicMessagesConfig so the block is kept, and
overridden to True on the providers that reach these transforms and reject the
block (Bedrock platform, Vertex, Azure for the chat path; Minimax, Azure, DeepSeek
for the messages path). Behavior for those providers is unchanged.

* Strip billing header on Bedrock invoke and Vertex messages pass-through

Two more subclasses reach the gated strip but inherited keep-by-default.
AmazonAnthropicClaudeConfig (Bedrock invoke) calls AnthropicConfig.transform_request,
which calls translate_system_message, and VertexAIPartnerModelsAnthropicMessagesConfig
(Vertex messages pass-through) calls super().transform_anthropic_messages_request.
Override should_strip_billing_metadata() to True on both.

Add a parametrized test asserting the flag for every first-party base (False) and
provider subclass (True), covering all overrides, plus a translate_system_message
regression test for the Bedrock invoke path.

* fix(cache): log hashed cache keys (#29890)

* fix(ui): save routing groups as list (#29889)

* Revert "fix(ui): save routing groups as list (#29889)" (#29928)

This reverts commit 9b1f78f.

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider (#29842)

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider

Registers parasail in the openai_like JSON provider loader with both
/v1/chat/completions and /v1/responses support. Parasail's Responses API
rejects store:true and any request that omits store, so the loader gains a
force_store_false special_handling flag; the parasail entry sets it and
the generated Responses config overrides store=false on every call. This
keeps callers from hitting "State storage not supported" and matches what
Parasail's docs require.

Adds the PARASAIL enum value, listing under openai_compatible_providers,
provider documentation at docs/my-website/docs/providers/parasail.md, and
a focused unit test file under tests/test_litellm/llms/parasail/ that
covers JSON registration, chat URL construction, Responses URL
construction with PARASAIL_API_BASE override, and the force_store_false
regression in both the caller-sent-store=true and caller-omitted cases.

* fix(parasail): register in provider_endpoints_support, drop in-repo docs

Greptile review feedback. The provider doc belongs in the litellm-docs
repo, not this one's docs/my-website tree; removing it here. Adds the
parasail entry to provider_endpoints_support.json so the
check_provider_folders_documented.py CI check passes (chat_completions
and responses true; others false).

* fix: normalize Anthropic passthrough server tool usage (#29827)

* test(anthropic): cover server_tool_use dict cost tracking

* fix: normalize Anthropic server tool usage

(cherry picked from commit 982f726)

* fix: keep server tool usage subscriptable

(cherry picked from commit 70280b9)

---------

Co-authored-by: Genmin <joey@joeyroth.com>

* fix(proxy): fix typo generic_role_mappoings -> generic_role_mappings in ui_sso.py (#29753)

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* feat(proxy): add disable_budget_reservation general setting (#27639) (#29493)

* feat(proxy): add disable_budget_reservation general setting (#27639)

* feat(proxy): register disable_budget_reservation in ConfigGeneralSettings (#27639)

* docs(proxy): document disable_budget_reservation concurrency tradeoff (#27639)

* ci: re-trigger flaky docker build (prisma generate ECONNRESET)

* fix(proxy): warn and document budget enforcement tradeoff when disable_budget_reservation is set (#27639)

* feat(gemini_tts): adding support to Gemini TTS languageCode parameters (#29623)

* Adding support to Gemini TTS Language Code parameters

* Mapping Gemini TTS languageCode param in Docstring

* Use snake_case for language_code input keyMapping Gemini TTS languageCode param in Docstring

* Restoring files modified under enterprise/litellm_enterprise due to lint/formatting checks

---------

Co-authored-by: João Garrido <joaogarrido@google.com>

* feat(guardrails): capture user and model metadata in CrowdStrike AIDR (#29517)

* fix(proxy): require OpenAI path segment for shared Azure Cognitive Services domains

Address Greptile review: the `*.cognitiveservices.azure.com` /
`*.openai.azure.com` domains are shared by every Azure Cognitive Service
(Speech, Vision, Language, ...), so a hostname-only substring match
misclassified non-OpenAI Azure traffic as OpenAI routes.

- Replace the substring host test with suffix matching (rejects look-alike
  domains like cognitiveservices.azure.com.attacker.example).
- Add `_is_openai_compatible_url` that requires an OpenAI-style path marker
  (`/openai/` or `/v1/`) on the shared Azure domains, and use it in
  PassThroughEndpointLogging.is_openai_route (previously hostname-only).
- Add negative tests for Azure Speech/Vision paths and look-alike domains.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* fix: support Responses input in Redis semantic cache (#29581)

* fix: support responses input in redis semantic cache

* test: cover redis semantic prompt extraction

* test: handle blank redis semantic text fallbacks

* chore: remove async cache dead statement

* test: cover redis semantic cache miss paths

* fix: filter sensitive cache lookup kwargs

* chore: rerun ci after huggingface rate limit

* chore(ui): regenerate dashboard API types (npm run gen:api)

Sync src/lib/http/schema.d.ts with the proxy OpenAPI spec: adds the
disable_budget_reservation general-settings field and picks up the
RateLimitError docstring reindent. Fixes the gen:api CI drift check.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* test(bedrock): assert empty additionalModelRequestFields is omitted

The Converse transformer now drops an empty additionalModelRequestFields
block instead of sending it as `{}`. Update test_bedrock_top_k_param so
models without top_k support (llama3) assert the key is absent rather than
equal to an empty dict.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Kent <72616338+kingdoooo@users.noreply.github.com>
Co-authored-by: codgician <15964984+codgician@users.noreply.github.com>
Co-authored-by: Praveen Ghuge <95286176+pghuge-cloudwiz@users.noreply.github.com>
Co-authored-by: Roi <roytev@gmail.com>
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Liam Scott <liam@uilliam.com>
Co-authored-by: abhay23-AI <abhaytrivedi22@gmail.com>
Co-authored-by: Ceder Dens <cederdens@gmail.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>
Co-authored-by: Kai Huang <kaihuang724@gmail.com>
Co-authored-by: rinto <54238243+ririnto@users.noreply.github.com>
Co-authored-by: Genmin <joey@joeyroth.com>
Co-authored-by: Arnav Bhilwariya <arnavbhilwariya0408@gmail.com>
Co-authored-by: Armaan Sandhu <74664101+Ar-maan05@users.noreply.github.com>
Co-authored-by: João Garrido <48538534+johngarrido@users.noreply.github.com>
Co-authored-by: João Garrido <joaogarrido@google.com>
Co-authored-by: Kenan Yildirim <kenan@kenany.me>
Co-authored-by: Dávid Balatoni <balcsida@gmail.com>
michaelxer pushed a commit to michaelxer/litellm that referenced this pull request Jun 17, 2026
* feat(bedrock_mantle): add SigV4/IAM auth to Responses API route (fixes BerriAI#29665) (BerriAI#29788)

* feat(responses): add default no-op sign_request to BaseResponsesAPIConfig

* feat(responses): call sign_request after body is final, send signed bytes when signed

* feat(bedrock_mantle): add SigV4 sign_request via composed BaseAWSLLM (bearer path)

* test(bedrock_mantle): cover SigV4 access-key, AssumeRole, body bytes, region/auth consistency

* feat(bedrock_mantle): defer auth to sign_request; validate_environment no longer requires bearer

* docs(bedrock_mantle): document SigV4 + Bearer auth on Responses route

* test(responses): cover fake-stream signing order and mantle bearer arg/env precedence

* fix(bedrock_mantle): wrap all botocore credential errors with both-paths guidance

* fix(bedrock_mantle): catch specific credential errors, not all BotoCoreError, so STS transport failures are not masked

* fix(bedrock_mantle): sign the compact Responses route too, not just create

* fix(github-copilot): route per-model on /v1/responses based on model info (BerriAI#29747)

* feat(focus): add GCS destination for FOCUS export (BerriAI#29751)

* test: add failing tests for FocusGCSDestination

* feat: add FocusGCSDestination reusing GCSBucketBase auth

* feat: register FocusGCSDestination in factory; export from __init__

* fix(focus): preserve GCS_PATH_SERVICE_ACCOUNT when service_account_json not in config

* style: apply Black formatting to gcs_destination and tests

* style: apply Black formatting to factory.py

* fix(bedrock): omit empty additionalModelRequestFields and system from Converse API payload (BerriAI#29565)

Amazon Nova Pro (and other strict Bedrock models) return 400 Malformed input
request when additionalModelRequestFields: {} or system: [] are present in the
payload. Both fields are optional in CommonRequestObject (total=False) and must
be omitted rather than sent as empty structures.

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible in pass-through cost tracking (BerriAI#29730)

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible

Azure OpenAI resources created via the newer "Azure AI Foundry" /
Cognitive Services pathway live on `*.cognitiveservices.azure.com`
subdomains, not the older `openai.azure.com`. Both are valid Azure
OpenAI surfaces in production today.

The OpenAI pass-through cost-tracking handler hard-codes only the older
hostname in five places (four `is_openai_*_route` methods on
OpenAIPassthroughLoggingHandler, plus is_openai_route on
PassThroughEndpointLogging). As a result, calls from newer Azure
deployments are silently classified as "not an OpenAI route", the
dispatch into the cost-tracking handler is skipped, and tokens/cost
never get extracted into LiteLLM_SpendLogs — the row gets written with
prompt_tokens=0, completion_tokens=0, spend=0, model='unknown'.

Reproduced 2026-06-04 against a real Azure OpenAI deployment on
`*.cognitiveservices.azure.com` proxied through LiteLLM v1.88.0.

Fix: factor the hostname check into a single helper
`_is_openai_compatible_host` listing all three recognized surfaces
(api.openai.com, openai.azure.com, cognitiveservices.azure.com), and
have all five call sites delegate to it. Purely additive — never
weakens recognition for the originally-supported hostnames.

Adds a test
`test_is_openai_route_recognizes_cognitiveservices_azure_com` that
exercises all four `is_openai_*_route` static methods against
`*.cognitiveservices.azure.com` URLs (positive cases per route + a
small cross-route negative to confirm route-specific path matching
still works on the new hostname).

Out of scope for this PR (separate followup):
  - `openai_passthrough_handler` calls chat/completions
    `transform_response` on Responses API payloads (`output:` not
    `choices:`), which throws inside the dispatch and drops the
    SpendLogs row entirely. Recognized + tracked separately.

* ci: trigger fresh run

Empty commit to re-run checks. The previous auth-and-jwt failure was
a transient HuggingFace Hub 429 rate-limit hitting tokenizer downloads
in tests/proxy_unit_tests/test_custom_tokenizer_bug.py — unrelated to
this PR's scope (hostname recognition in pass-through cost tracking).
No code change.

---------

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* fix(responses): preserve forced-function tool_choice name in Responses to Chat transform (BerriAI#29812)

The Responses API forces a specific function with a top-level name
({"type": "function", "name": "X"}), but _transform_tool_choice only handled the
nested Chat Completions shape and fell through to returning "required" for the flat
form, silently dropping the function name and degrading a forced function call to
force-any-tool. Map the flat Responses shape to the nested Chat shape, keeping the
"required" fallback when no name is present.

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic (BerriAI#29584)

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic

PR BerriAI#20951 strips system blocks beginning with "x-anthropic-billing-header:" for
every Anthropic target. That block is how the first-party Anthropic API recognizes
Claude Code subscription (OAuth) traffic, so dropping it makes requests that carry
only that block, such as the auto-mode tool-safety classifier, fail with a
misleading 429 rate_limit_error; normal turns still work because they also carry
the "You are Claude Code" identity block.

Gate the strip behind should_strip_billing_metadata(), defaulting to False on the
first-party AnthropicConfig and AnthropicMessagesConfig so the block is kept, and
overridden to True on the providers that reach these transforms and reject the
block (Bedrock platform, Vertex, Azure for the chat path; Minimax, Azure, DeepSeek
for the messages path). Behavior for those providers is unchanged.

* Strip billing header on Bedrock invoke and Vertex messages pass-through

Two more subclasses reach the gated strip but inherited keep-by-default.
AmazonAnthropicClaudeConfig (Bedrock invoke) calls AnthropicConfig.transform_request,
which calls translate_system_message, and VertexAIPartnerModelsAnthropicMessagesConfig
(Vertex messages pass-through) calls super().transform_anthropic_messages_request.
Override should_strip_billing_metadata() to True on both.

Add a parametrized test asserting the flag for every first-party base (False) and
provider subclass (True), covering all overrides, plus a translate_system_message
regression test for the Bedrock invoke path.

* fix(cache): log hashed cache keys (BerriAI#29890)

* fix(ui): save routing groups as list (BerriAI#29889)

* Revert "fix(ui): save routing groups as list (BerriAI#29889)" (BerriAI#29928)

This reverts commit 9b1f78f.

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider (BerriAI#29842)

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider

Registers parasail in the openai_like JSON provider loader with both
/v1/chat/completions and /v1/responses support. Parasail's Responses API
rejects store:true and any request that omits store, so the loader gains a
force_store_false special_handling flag; the parasail entry sets it and
the generated Responses config overrides store=false on every call. This
keeps callers from hitting "State storage not supported" and matches what
Parasail's docs require.

Adds the PARASAIL enum value, listing under openai_compatible_providers,
provider documentation at docs/my-website/docs/providers/parasail.md, and
a focused unit test file under tests/test_litellm/llms/parasail/ that
covers JSON registration, chat URL construction, Responses URL
construction with PARASAIL_API_BASE override, and the force_store_false
regression in both the caller-sent-store=true and caller-omitted cases.

* fix(parasail): register in provider_endpoints_support, drop in-repo docs

Greptile review feedback. The provider doc belongs in the litellm-docs
repo, not this one's docs/my-website tree; removing it here. Adds the
parasail entry to provider_endpoints_support.json so the
check_provider_folders_documented.py CI check passes (chat_completions
and responses true; others false).

* fix: normalize Anthropic passthrough server tool usage (BerriAI#29827)

* test(anthropic): cover server_tool_use dict cost tracking

* fix: normalize Anthropic server tool usage

(cherry picked from commit 982f726)

* fix: keep server tool usage subscriptable

(cherry picked from commit 70280b9)

---------

Co-authored-by: Genmin <joey@joeyroth.com>

* fix(proxy): fix typo generic_role_mappoings -> generic_role_mappings in ui_sso.py (BerriAI#29753)

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639) (BerriAI#29493)

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639)

* feat(proxy): register disable_budget_reservation in ConfigGeneralSettings (BerriAI#27639)

* docs(proxy): document disable_budget_reservation concurrency tradeoff (BerriAI#27639)

* ci: re-trigger flaky docker build (prisma generate ECONNRESET)

* fix(proxy): warn and document budget enforcement tradeoff when disable_budget_reservation is set (BerriAI#27639)

* feat(gemini_tts): adding support to Gemini TTS languageCode parameters (BerriAI#29623)

* Adding support to Gemini TTS Language Code parameters

* Mapping Gemini TTS languageCode param in Docstring

* Use snake_case for language_code input keyMapping Gemini TTS languageCode param in Docstring

* Restoring files modified under enterprise/litellm_enterprise due to lint/formatting checks

---------

Co-authored-by: João Garrido <joaogarrido@google.com>

* feat(guardrails): capture user and model metadata in CrowdStrike AIDR (BerriAI#29517)

* fix(proxy): require OpenAI path segment for shared Azure Cognitive Services domains

Address Greptile review: the `*.cognitiveservices.azure.com` /
`*.openai.azure.com` domains are shared by every Azure Cognitive Service
(Speech, Vision, Language, ...), so a hostname-only substring match
misclassified non-OpenAI Azure traffic as OpenAI routes.

- Replace the substring host test with suffix matching (rejects look-alike
  domains like cognitiveservices.azure.com.attacker.example).
- Add `_is_openai_compatible_url` that requires an OpenAI-style path marker
  (`/openai/` or `/v1/`) on the shared Azure domains, and use it in
  PassThroughEndpointLogging.is_openai_route (previously hostname-only).
- Add negative tests for Azure Speech/Vision paths and look-alike domains.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* fix: support Responses input in Redis semantic cache (BerriAI#29581)

* fix: support responses input in redis semantic cache

* test: cover redis semantic prompt extraction

* test: handle blank redis semantic text fallbacks

* chore: remove async cache dead statement

* test: cover redis semantic cache miss paths

* fix: filter sensitive cache lookup kwargs

* chore: rerun ci after huggingface rate limit

* chore(ui): regenerate dashboard API types (npm run gen:api)

Sync src/lib/http/schema.d.ts with the proxy OpenAPI spec: adds the
disable_budget_reservation general-settings field and picks up the
RateLimitError docstring reindent. Fixes the gen:api CI drift check.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* test(bedrock): assert empty additionalModelRequestFields is omitted

The Converse transformer now drops an empty additionalModelRequestFields
block instead of sending it as `{}`. Update test_bedrock_top_k_param so
models without top_k support (llama3) assert the key is absent rather than
equal to an empty dict.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Kent <72616338+kingdoooo@users.noreply.github.com>
Co-authored-by: codgician <15964984+codgician@users.noreply.github.com>
Co-authored-by: Praveen Ghuge <95286176+pghuge-cloudwiz@users.noreply.github.com>
Co-authored-by: Roi <roytev@gmail.com>
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Liam Scott <liam@uilliam.com>
Co-authored-by: abhay23-AI <abhaytrivedi22@gmail.com>
Co-authored-by: Ceder Dens <cederdens@gmail.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>
Co-authored-by: Kai Huang <kaihuang724@gmail.com>
Co-authored-by: rinto <54238243+ririnto@users.noreply.github.com>
Co-authored-by: Genmin <joey@joeyroth.com>
Co-authored-by: Arnav Bhilwariya <arnavbhilwariya0408@gmail.com>
Co-authored-by: Armaan Sandhu <74664101+Ar-maan05@users.noreply.github.com>
Co-authored-by: João Garrido <48538534+johngarrido@users.noreply.github.com>
Co-authored-by: João Garrido <joaogarrido@google.com>
Co-authored-by: Kenan Yildirim <kenan@kenany.me>
Co-authored-by: Dávid Balatoni <balcsida@gmail.com>
michaelxer pushed a commit to michaelxer/litellm that referenced this pull request Jun 17, 2026
* feat(bedrock_mantle): add SigV4/IAM auth to Responses API route (fixes BerriAI#29665) (BerriAI#29788)

* feat(responses): add default no-op sign_request to BaseResponsesAPIConfig

* feat(responses): call sign_request after body is final, send signed bytes when signed

* feat(bedrock_mantle): add SigV4 sign_request via composed BaseAWSLLM (bearer path)

* test(bedrock_mantle): cover SigV4 access-key, AssumeRole, body bytes, region/auth consistency

* feat(bedrock_mantle): defer auth to sign_request; validate_environment no longer requires bearer

* docs(bedrock_mantle): document SigV4 + Bearer auth on Responses route

* test(responses): cover fake-stream signing order and mantle bearer arg/env precedence

* fix(bedrock_mantle): wrap all botocore credential errors with both-paths guidance

* fix(bedrock_mantle): catch specific credential errors, not all BotoCoreError, so STS transport failures are not masked

* fix(bedrock_mantle): sign the compact Responses route too, not just create

* fix(github-copilot): route per-model on /v1/responses based on model info (BerriAI#29747)

* feat(focus): add GCS destination for FOCUS export (BerriAI#29751)

* test: add failing tests for FocusGCSDestination

* feat: add FocusGCSDestination reusing GCSBucketBase auth

* feat: register FocusGCSDestination in factory; export from __init__

* fix(focus): preserve GCS_PATH_SERVICE_ACCOUNT when service_account_json not in config

* style: apply Black formatting to gcs_destination and tests

* style: apply Black formatting to factory.py

* fix(bedrock): omit empty additionalModelRequestFields and system from Converse API payload (BerriAI#29565)

Amazon Nova Pro (and other strict Bedrock models) return 400 Malformed input
request when additionalModelRequestFields: {} or system: [] are present in the
payload. Both fields are optional in CommonRequestObject (total=False) and must
be omitted rather than sent as empty structures.

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible in pass-through cost tracking (BerriAI#29730)

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible

Azure OpenAI resources created via the newer "Azure AI Foundry" /
Cognitive Services pathway live on `*.cognitiveservices.azure.com`
subdomains, not the older `openai.azure.com`. Both are valid Azure
OpenAI surfaces in production today.

The OpenAI pass-through cost-tracking handler hard-codes only the older
hostname in five places (four `is_openai_*_route` methods on
OpenAIPassthroughLoggingHandler, plus is_openai_route on
PassThroughEndpointLogging). As a result, calls from newer Azure
deployments are silently classified as "not an OpenAI route", the
dispatch into the cost-tracking handler is skipped, and tokens/cost
never get extracted into LiteLLM_SpendLogs — the row gets written with
prompt_tokens=0, completion_tokens=0, spend=0, model='unknown'.

Reproduced 2026-06-04 against a real Azure OpenAI deployment on
`*.cognitiveservices.azure.com` proxied through LiteLLM v1.88.0.

Fix: factor the hostname check into a single helper
`_is_openai_compatible_host` listing all three recognized surfaces
(api.openai.com, openai.azure.com, cognitiveservices.azure.com), and
have all five call sites delegate to it. Purely additive — never
weakens recognition for the originally-supported hostnames.

Adds a test
`test_is_openai_route_recognizes_cognitiveservices_azure_com` that
exercises all four `is_openai_*_route` static methods against
`*.cognitiveservices.azure.com` URLs (positive cases per route + a
small cross-route negative to confirm route-specific path matching
still works on the new hostname).

Out of scope for this PR (separate followup):
  - `openai_passthrough_handler` calls chat/completions
    `transform_response` on Responses API payloads (`output:` not
    `choices:`), which throws inside the dispatch and drops the
    SpendLogs row entirely. Recognized + tracked separately.

* ci: trigger fresh run

Empty commit to re-run checks. The previous auth-and-jwt failure was
a transient HuggingFace Hub 429 rate-limit hitting tokenizer downloads
in tests/proxy_unit_tests/test_custom_tokenizer_bug.py — unrelated to
this PR's scope (hostname recognition in pass-through cost tracking).
No code change.

---------

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* fix(responses): preserve forced-function tool_choice name in Responses to Chat transform (BerriAI#29812)

The Responses API forces a specific function with a top-level name
({"type": "function", "name": "X"}), but _transform_tool_choice only handled the
nested Chat Completions shape and fell through to returning "required" for the flat
form, silently dropping the function name and degrading a forced function call to
force-any-tool. Map the flat Responses shape to the nested Chat shape, keeping the
"required" fallback when no name is present.

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic (BerriAI#29584)

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic

PR BerriAI#20951 strips system blocks beginning with "x-anthropic-billing-header:" for
every Anthropic target. That block is how the first-party Anthropic API recognizes
Claude Code subscription (OAuth) traffic, so dropping it makes requests that carry
only that block, such as the auto-mode tool-safety classifier, fail with a
misleading 429 rate_limit_error; normal turns still work because they also carry
the "You are Claude Code" identity block.

Gate the strip behind should_strip_billing_metadata(), defaulting to False on the
first-party AnthropicConfig and AnthropicMessagesConfig so the block is kept, and
overridden to True on the providers that reach these transforms and reject the
block (Bedrock platform, Vertex, Azure for the chat path; Minimax, Azure, DeepSeek
for the messages path). Behavior for those providers is unchanged.

* Strip billing header on Bedrock invoke and Vertex messages pass-through

Two more subclasses reach the gated strip but inherited keep-by-default.
AmazonAnthropicClaudeConfig (Bedrock invoke) calls AnthropicConfig.transform_request,
which calls translate_system_message, and VertexAIPartnerModelsAnthropicMessagesConfig
(Vertex messages pass-through) calls super().transform_anthropic_messages_request.
Override should_strip_billing_metadata() to True on both.

Add a parametrized test asserting the flag for every first-party base (False) and
provider subclass (True), covering all overrides, plus a translate_system_message
regression test for the Bedrock invoke path.

* fix(cache): log hashed cache keys (BerriAI#29890)

* fix(ui): save routing groups as list (BerriAI#29889)

* Revert "fix(ui): save routing groups as list (BerriAI#29889)" (BerriAI#29928)

This reverts commit 9b1f78f.

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider (BerriAI#29842)

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider

Registers parasail in the openai_like JSON provider loader with both
/v1/chat/completions and /v1/responses support. Parasail's Responses API
rejects store:true and any request that omits store, so the loader gains a
force_store_false special_handling flag; the parasail entry sets it and
the generated Responses config overrides store=false on every call. This
keeps callers from hitting "State storage not supported" and matches what
Parasail's docs require.

Adds the PARASAIL enum value, listing under openai_compatible_providers,
provider documentation at docs/my-website/docs/providers/parasail.md, and
a focused unit test file under tests/test_litellm/llms/parasail/ that
covers JSON registration, chat URL construction, Responses URL
construction with PARASAIL_API_BASE override, and the force_store_false
regression in both the caller-sent-store=true and caller-omitted cases.

* fix(parasail): register in provider_endpoints_support, drop in-repo docs

Greptile review feedback. The provider doc belongs in the litellm-docs
repo, not this one's docs/my-website tree; removing it here. Adds the
parasail entry to provider_endpoints_support.json so the
check_provider_folders_documented.py CI check passes (chat_completions
and responses true; others false).

* fix: normalize Anthropic passthrough server tool usage (BerriAI#29827)

* test(anthropic): cover server_tool_use dict cost tracking

* fix: normalize Anthropic server tool usage

(cherry picked from commit 982f726)

* fix: keep server tool usage subscriptable

(cherry picked from commit 70280b9)

---------

Co-authored-by: Genmin <joey@joeyroth.com>

* fix(proxy): fix typo generic_role_mappoings -> generic_role_mappings in ui_sso.py (BerriAI#29753)

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639) (BerriAI#29493)

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639)

* feat(proxy): register disable_budget_reservation in ConfigGeneralSettings (BerriAI#27639)

* docs(proxy): document disable_budget_reservation concurrency tradeoff (BerriAI#27639)

* ci: re-trigger flaky docker build (prisma generate ECONNRESET)

* fix(proxy): warn and document budget enforcement tradeoff when disable_budget_reservation is set (BerriAI#27639)

* feat(gemini_tts): adding support to Gemini TTS languageCode parameters (BerriAI#29623)

* Adding support to Gemini TTS Language Code parameters

* Mapping Gemini TTS languageCode param in Docstring

* Use snake_case for language_code input keyMapping Gemini TTS languageCode param in Docstring

* Restoring files modified under enterprise/litellm_enterprise due to lint/formatting checks

---------

Co-authored-by: João Garrido <joaogarrido@google.com>

* feat(guardrails): capture user and model metadata in CrowdStrike AIDR (BerriAI#29517)

* fix(proxy): require OpenAI path segment for shared Azure Cognitive Services domains

Address Greptile review: the `*.cognitiveservices.azure.com` /
`*.openai.azure.com` domains are shared by every Azure Cognitive Service
(Speech, Vision, Language, ...), so a hostname-only substring match
misclassified non-OpenAI Azure traffic as OpenAI routes.

- Replace the substring host test with suffix matching (rejects look-alike
  domains like cognitiveservices.azure.com.attacker.example).
- Add `_is_openai_compatible_url` that requires an OpenAI-style path marker
  (`/openai/` or `/v1/`) on the shared Azure domains, and use it in
  PassThroughEndpointLogging.is_openai_route (previously hostname-only).
- Add negative tests for Azure Speech/Vision paths and look-alike domains.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* fix: support Responses input in Redis semantic cache (BerriAI#29581)

* fix: support responses input in redis semantic cache

* test: cover redis semantic prompt extraction

* test: handle blank redis semantic text fallbacks

* chore: remove async cache dead statement

* test: cover redis semantic cache miss paths

* fix: filter sensitive cache lookup kwargs

* chore: rerun ci after huggingface rate limit

* chore(ui): regenerate dashboard API types (npm run gen:api)

Sync src/lib/http/schema.d.ts with the proxy OpenAPI spec: adds the
disable_budget_reservation general-settings field and picks up the
RateLimitError docstring reindent. Fixes the gen:api CI drift check.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* test(bedrock): assert empty additionalModelRequestFields is omitted

The Converse transformer now drops an empty additionalModelRequestFields
block instead of sending it as `{}`. Update test_bedrock_top_k_param so
models without top_k support (llama3) assert the key is absent rather than
equal to an empty dict.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Kent <72616338+kingdoooo@users.noreply.github.com>
Co-authored-by: codgician <15964984+codgician@users.noreply.github.com>
Co-authored-by: Praveen Ghuge <95286176+pghuge-cloudwiz@users.noreply.github.com>
Co-authored-by: Roi <roytev@gmail.com>
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Liam Scott <liam@uilliam.com>
Co-authored-by: abhay23-AI <abhaytrivedi22@gmail.com>
Co-authored-by: Ceder Dens <cederdens@gmail.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>
Co-authored-by: Kai Huang <kaihuang724@gmail.com>
Co-authored-by: rinto <54238243+ririnto@users.noreply.github.com>
Co-authored-by: Genmin <joey@joeyroth.com>
Co-authored-by: Arnav Bhilwariya <arnavbhilwariya0408@gmail.com>
Co-authored-by: Armaan Sandhu <74664101+Ar-maan05@users.noreply.github.com>
Co-authored-by: João Garrido <48538534+johngarrido@users.noreply.github.com>
Co-authored-by: João Garrido <joaogarrido@google.com>
Co-authored-by: Kenan Yildirim <kenan@kenany.me>
Co-authored-by: Dávid Balatoni <balcsida@gmail.com>
koladefaj pushed a commit to koladefaj/litellm that referenced this pull request Jun 17, 2026
* feat(bedrock_mantle): add SigV4/IAM auth to Responses API route (fixes BerriAI#29665) (BerriAI#29788)

* feat(responses): add default no-op sign_request to BaseResponsesAPIConfig

* feat(responses): call sign_request after body is final, send signed bytes when signed

* feat(bedrock_mantle): add SigV4 sign_request via composed BaseAWSLLM (bearer path)

* test(bedrock_mantle): cover SigV4 access-key, AssumeRole, body bytes, region/auth consistency

* feat(bedrock_mantle): defer auth to sign_request; validate_environment no longer requires bearer

* docs(bedrock_mantle): document SigV4 + Bearer auth on Responses route

* test(responses): cover fake-stream signing order and mantle bearer arg/env precedence

* fix(bedrock_mantle): wrap all botocore credential errors with both-paths guidance

* fix(bedrock_mantle): catch specific credential errors, not all BotoCoreError, so STS transport failures are not masked

* fix(bedrock_mantle): sign the compact Responses route too, not just create

* fix(github-copilot): route per-model on /v1/responses based on model info (BerriAI#29747)

* feat(focus): add GCS destination for FOCUS export (BerriAI#29751)

* test: add failing tests for FocusGCSDestination

* feat: add FocusGCSDestination reusing GCSBucketBase auth

* feat: register FocusGCSDestination in factory; export from __init__

* fix(focus): preserve GCS_PATH_SERVICE_ACCOUNT when service_account_json not in config

* style: apply Black formatting to gcs_destination and tests

* style: apply Black formatting to factory.py

* fix(bedrock): omit empty additionalModelRequestFields and system from Converse API payload (BerriAI#29565)

Amazon Nova Pro (and other strict Bedrock models) return 400 Malformed input
request when additionalModelRequestFields: {} or system: [] are present in the
payload. Both fields are optional in CommonRequestObject (total=False) and must
be omitted rather than sent as empty structures.

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible in pass-through cost tracking (BerriAI#29730)

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible

Azure OpenAI resources created via the newer "Azure AI Foundry" /
Cognitive Services pathway live on `*.cognitiveservices.azure.com`
subdomains, not the older `openai.azure.com`. Both are valid Azure
OpenAI surfaces in production today.

The OpenAI pass-through cost-tracking handler hard-codes only the older
hostname in five places (four `is_openai_*_route` methods on
OpenAIPassthroughLoggingHandler, plus is_openai_route on
PassThroughEndpointLogging). As a result, calls from newer Azure
deployments are silently classified as "not an OpenAI route", the
dispatch into the cost-tracking handler is skipped, and tokens/cost
never get extracted into LiteLLM_SpendLogs — the row gets written with
prompt_tokens=0, completion_tokens=0, spend=0, model='unknown'.

Reproduced 2026-06-04 against a real Azure OpenAI deployment on
`*.cognitiveservices.azure.com` proxied through LiteLLM v1.88.0.

Fix: factor the hostname check into a single helper
`_is_openai_compatible_host` listing all three recognized surfaces
(api.openai.com, openai.azure.com, cognitiveservices.azure.com), and
have all five call sites delegate to it. Purely additive — never
weakens recognition for the originally-supported hostnames.

Adds a test
`test_is_openai_route_recognizes_cognitiveservices_azure_com` that
exercises all four `is_openai_*_route` static methods against
`*.cognitiveservices.azure.com` URLs (positive cases per route + a
small cross-route negative to confirm route-specific path matching
still works on the new hostname).

Out of scope for this PR (separate followup):
  - `openai_passthrough_handler` calls chat/completions
    `transform_response` on Responses API payloads (`output:` not
    `choices:`), which throws inside the dispatch and drops the
    SpendLogs row entirely. Recognized + tracked separately.

* ci: trigger fresh run

Empty commit to re-run checks. The previous auth-and-jwt failure was
a transient HuggingFace Hub 429 rate-limit hitting tokenizer downloads
in tests/proxy_unit_tests/test_custom_tokenizer_bug.py — unrelated to
this PR's scope (hostname recognition in pass-through cost tracking).
No code change.

---------

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* fix(responses): preserve forced-function tool_choice name in Responses to Chat transform (BerriAI#29812)

The Responses API forces a specific function with a top-level name
({"type": "function", "name": "X"}), but _transform_tool_choice only handled the
nested Chat Completions shape and fell through to returning "required" for the flat
form, silently dropping the function name and degrading a forced function call to
force-any-tool. Map the flat Responses shape to the nested Chat shape, keeping the
"required" fallback when no name is present.

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic (BerriAI#29584)

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic

PR BerriAI#20951 strips system blocks beginning with "x-anthropic-billing-header:" for
every Anthropic target. That block is how the first-party Anthropic API recognizes
Claude Code subscription (OAuth) traffic, so dropping it makes requests that carry
only that block, such as the auto-mode tool-safety classifier, fail with a
misleading 429 rate_limit_error; normal turns still work because they also carry
the "You are Claude Code" identity block.

Gate the strip behind should_strip_billing_metadata(), defaulting to False on the
first-party AnthropicConfig and AnthropicMessagesConfig so the block is kept, and
overridden to True on the providers that reach these transforms and reject the
block (Bedrock platform, Vertex, Azure for the chat path; Minimax, Azure, DeepSeek
for the messages path). Behavior for those providers is unchanged.

* Strip billing header on Bedrock invoke and Vertex messages pass-through

Two more subclasses reach the gated strip but inherited keep-by-default.
AmazonAnthropicClaudeConfig (Bedrock invoke) calls AnthropicConfig.transform_request,
which calls translate_system_message, and VertexAIPartnerModelsAnthropicMessagesConfig
(Vertex messages pass-through) calls super().transform_anthropic_messages_request.
Override should_strip_billing_metadata() to True on both.

Add a parametrized test asserting the flag for every first-party base (False) and
provider subclass (True), covering all overrides, plus a translate_system_message
regression test for the Bedrock invoke path.

* fix(cache): log hashed cache keys (BerriAI#29890)

* fix(ui): save routing groups as list (BerriAI#29889)

* Revert "fix(ui): save routing groups as list (BerriAI#29889)" (BerriAI#29928)

This reverts commit 9b1f78f.

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider (BerriAI#29842)

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider

Registers parasail in the openai_like JSON provider loader with both
/v1/chat/completions and /v1/responses support. Parasail's Responses API
rejects store:true and any request that omits store, so the loader gains a
force_store_false special_handling flag; the parasail entry sets it and
the generated Responses config overrides store=false on every call. This
keeps callers from hitting "State storage not supported" and matches what
Parasail's docs require.

Adds the PARASAIL enum value, listing under openai_compatible_providers,
provider documentation at docs/my-website/docs/providers/parasail.md, and
a focused unit test file under tests/test_litellm/llms/parasail/ that
covers JSON registration, chat URL construction, Responses URL
construction with PARASAIL_API_BASE override, and the force_store_false
regression in both the caller-sent-store=true and caller-omitted cases.

* fix(parasail): register in provider_endpoints_support, drop in-repo docs

Greptile review feedback. The provider doc belongs in the litellm-docs
repo, not this one's docs/my-website tree; removing it here. Adds the
parasail entry to provider_endpoints_support.json so the
check_provider_folders_documented.py CI check passes (chat_completions
and responses true; others false).

* fix: normalize Anthropic passthrough server tool usage (BerriAI#29827)

* test(anthropic): cover server_tool_use dict cost tracking

* fix: normalize Anthropic server tool usage

(cherry picked from commit 982f726)

* fix: keep server tool usage subscriptable

(cherry picked from commit 70280b9)

---------

Co-authored-by: Genmin <joey@joeyroth.com>

* fix(proxy): fix typo generic_role_mappoings -> generic_role_mappings in ui_sso.py (BerriAI#29753)

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639) (BerriAI#29493)

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639)

* feat(proxy): register disable_budget_reservation in ConfigGeneralSettings (BerriAI#27639)

* docs(proxy): document disable_budget_reservation concurrency tradeoff (BerriAI#27639)

* ci: re-trigger flaky docker build (prisma generate ECONNRESET)

* fix(proxy): warn and document budget enforcement tradeoff when disable_budget_reservation is set (BerriAI#27639)

* feat(gemini_tts): adding support to Gemini TTS languageCode parameters (BerriAI#29623)

* Adding support to Gemini TTS Language Code parameters

* Mapping Gemini TTS languageCode param in Docstring

* Use snake_case for language_code input keyMapping Gemini TTS languageCode param in Docstring

* Restoring files modified under enterprise/litellm_enterprise due to lint/formatting checks

---------

Co-authored-by: João Garrido <joaogarrido@google.com>

* feat(guardrails): capture user and model metadata in CrowdStrike AIDR (BerriAI#29517)

* fix(proxy): require OpenAI path segment for shared Azure Cognitive Services domains

Address Greptile review: the `*.cognitiveservices.azure.com` /
`*.openai.azure.com` domains are shared by every Azure Cognitive Service
(Speech, Vision, Language, ...), so a hostname-only substring match
misclassified non-OpenAI Azure traffic as OpenAI routes.

- Replace the substring host test with suffix matching (rejects look-alike
  domains like cognitiveservices.azure.com.attacker.example).
- Add `_is_openai_compatible_url` that requires an OpenAI-style path marker
  (`/openai/` or `/v1/`) on the shared Azure domains, and use it in
  PassThroughEndpointLogging.is_openai_route (previously hostname-only).
- Add negative tests for Azure Speech/Vision paths and look-alike domains.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* fix: support Responses input in Redis semantic cache (BerriAI#29581)

* fix: support responses input in redis semantic cache

* test: cover redis semantic prompt extraction

* test: handle blank redis semantic text fallbacks

* chore: remove async cache dead statement

* test: cover redis semantic cache miss paths

* fix: filter sensitive cache lookup kwargs

* chore: rerun ci after huggingface rate limit

* chore(ui): regenerate dashboard API types (npm run gen:api)

Sync src/lib/http/schema.d.ts with the proxy OpenAPI spec: adds the
disable_budget_reservation general-settings field and picks up the
RateLimitError docstring reindent. Fixes the gen:api CI drift check.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* test(bedrock): assert empty additionalModelRequestFields is omitted

The Converse transformer now drops an empty additionalModelRequestFields
block instead of sending it as `{}`. Update test_bedrock_top_k_param so
models without top_k support (llama3) assert the key is absent rather than
equal to an empty dict.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Kent <72616338+kingdoooo@users.noreply.github.com>
Co-authored-by: codgician <15964984+codgician@users.noreply.github.com>
Co-authored-by: Praveen Ghuge <95286176+pghuge-cloudwiz@users.noreply.github.com>
Co-authored-by: Roi <roytev@gmail.com>
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Liam Scott <liam@uilliam.com>
Co-authored-by: abhay23-AI <abhaytrivedi22@gmail.com>
Co-authored-by: Ceder Dens <cederdens@gmail.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>
Co-authored-by: Kai Huang <kaihuang724@gmail.com>
Co-authored-by: rinto <54238243+ririnto@users.noreply.github.com>
Co-authored-by: Genmin <joey@joeyroth.com>
Co-authored-by: Arnav Bhilwariya <arnavbhilwariya0408@gmail.com>
Co-authored-by: Armaan Sandhu <74664101+Ar-maan05@users.noreply.github.com>
Co-authored-by: João Garrido <48538534+johngarrido@users.noreply.github.com>
Co-authored-by: João Garrido <joaogarrido@google.com>
Co-authored-by: Kenan Yildirim <kenan@kenany.me>
Co-authored-by: Dávid Balatoni <balcsida@gmail.com>
factnn pushed a commit to factnn/litellm that referenced this pull request Jun 18, 2026
* feat(bedrock_mantle): add SigV4/IAM auth to Responses API route (fixes BerriAI#29665) (BerriAI#29788)

* feat(responses): add default no-op sign_request to BaseResponsesAPIConfig

* feat(responses): call sign_request after body is final, send signed bytes when signed

* feat(bedrock_mantle): add SigV4 sign_request via composed BaseAWSLLM (bearer path)

* test(bedrock_mantle): cover SigV4 access-key, AssumeRole, body bytes, region/auth consistency

* feat(bedrock_mantle): defer auth to sign_request; validate_environment no longer requires bearer

* docs(bedrock_mantle): document SigV4 + Bearer auth on Responses route

* test(responses): cover fake-stream signing order and mantle bearer arg/env precedence

* fix(bedrock_mantle): wrap all botocore credential errors with both-paths guidance

* fix(bedrock_mantle): catch specific credential errors, not all BotoCoreError, so STS transport failures are not masked

* fix(bedrock_mantle): sign the compact Responses route too, not just create

* fix(github-copilot): route per-model on /v1/responses based on model info (BerriAI#29747)

* feat(focus): add GCS destination for FOCUS export (BerriAI#29751)

* test: add failing tests for FocusGCSDestination

* feat: add FocusGCSDestination reusing GCSBucketBase auth

* feat: register FocusGCSDestination in factory; export from __init__

* fix(focus): preserve GCS_PATH_SERVICE_ACCOUNT when service_account_json not in config

* style: apply Black formatting to gcs_destination and tests

* style: apply Black formatting to factory.py

* fix(bedrock): omit empty additionalModelRequestFields and system from Converse API payload (BerriAI#29565)

Amazon Nova Pro (and other strict Bedrock models) return 400 Malformed input
request when additionalModelRequestFields: {} or system: [] are present in the
payload. Both fields are optional in CommonRequestObject (total=False) and must
be omitted rather than sent as empty structures.

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible in pass-through cost tracking (BerriAI#29730)

* fix(proxy): recognize *.cognitiveservices.azure.com as OpenAI-compatible

Azure OpenAI resources created via the newer "Azure AI Foundry" /
Cognitive Services pathway live on `*.cognitiveservices.azure.com`
subdomains, not the older `openai.azure.com`. Both are valid Azure
OpenAI surfaces in production today.

The OpenAI pass-through cost-tracking handler hard-codes only the older
hostname in five places (four `is_openai_*_route` methods on
OpenAIPassthroughLoggingHandler, plus is_openai_route on
PassThroughEndpointLogging). As a result, calls from newer Azure
deployments are silently classified as "not an OpenAI route", the
dispatch into the cost-tracking handler is skipped, and tokens/cost
never get extracted into LiteLLM_SpendLogs — the row gets written with
prompt_tokens=0, completion_tokens=0, spend=0, model='unknown'.

Reproduced 2026-06-04 against a real Azure OpenAI deployment on
`*.cognitiveservices.azure.com` proxied through LiteLLM v1.88.0.

Fix: factor the hostname check into a single helper
`_is_openai_compatible_host` listing all three recognized surfaces
(api.openai.com, openai.azure.com, cognitiveservices.azure.com), and
have all five call sites delegate to it. Purely additive — never
weakens recognition for the originally-supported hostnames.

Adds a test
`test_is_openai_route_recognizes_cognitiveservices_azure_com` that
exercises all four `is_openai_*_route` static methods against
`*.cognitiveservices.azure.com` URLs (positive cases per route + a
small cross-route negative to confirm route-specific path matching
still works on the new hostname).

Out of scope for this PR (separate followup):
  - `openai_passthrough_handler` calls chat/completions
    `transform_response` on Responses API payloads (`output:` not
    `choices:`), which throws inside the dispatch and drops the
    SpendLogs row entirely. Recognized + tracked separately.

* ci: trigger fresh run

Empty commit to re-run checks. The previous auth-and-jwt failure was
a transient HuggingFace Hub 429 rate-limit hitting tokenizer downloads
in tests/proxy_unit_tests/test_custom_tokenizer_bug.py — unrelated to
this PR's scope (hostname recognition in pass-through cost tracking).
No code change.

---------

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* fix(responses): preserve forced-function tool_choice name in Responses to Chat transform (BerriAI#29812)

The Responses API forces a specific function with a top-level name
({"type": "function", "name": "X"}), but _transform_tool_choice only handled the
nested Chat Completions shape and fell through to returning "required" for the flat
form, silently dropping the function name and degrading a forced function call to
force-any-tool. Map the flat Responses shape to the nested Chat shape, keeping the
"required" fallback when no name is present.

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic (BerriAI#29584)

* Preserve x-anthropic-billing-header system blocks for first-party Anthropic

PR BerriAI#20951 strips system blocks beginning with "x-anthropic-billing-header:" for
every Anthropic target. That block is how the first-party Anthropic API recognizes
Claude Code subscription (OAuth) traffic, so dropping it makes requests that carry
only that block, such as the auto-mode tool-safety classifier, fail with a
misleading 429 rate_limit_error; normal turns still work because they also carry
the "You are Claude Code" identity block.

Gate the strip behind should_strip_billing_metadata(), defaulting to False on the
first-party AnthropicConfig and AnthropicMessagesConfig so the block is kept, and
overridden to True on the providers that reach these transforms and reject the
block (Bedrock platform, Vertex, Azure for the chat path; Minimax, Azure, DeepSeek
for the messages path). Behavior for those providers is unchanged.

* Strip billing header on Bedrock invoke and Vertex messages pass-through

Two more subclasses reach the gated strip but inherited keep-by-default.
AmazonAnthropicClaudeConfig (Bedrock invoke) calls AnthropicConfig.transform_request,
which calls translate_system_message, and VertexAIPartnerModelsAnthropicMessagesConfig
(Vertex messages pass-through) calls super().transform_anthropic_messages_request.
Override should_strip_billing_metadata() to True on both.

Add a parametrized test asserting the flag for every first-party base (False) and
provider subclass (True), covering all overrides, plus a translate_system_message
regression test for the Bedrock invoke path.

* fix(cache): log hashed cache keys (BerriAI#29890)

* fix(ui): save routing groups as list (BerriAI#29889)

* Revert "fix(ui): save routing groups as list (BerriAI#29889)" (BerriAI#29928)

This reverts commit 9b1f78f.

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider (BerriAI#29842)

* feat(parasail): add Parasail as a JSON-configured OpenAI-compatible provider

Registers parasail in the openai_like JSON provider loader with both
/v1/chat/completions and /v1/responses support. Parasail's Responses API
rejects store:true and any request that omits store, so the loader gains a
force_store_false special_handling flag; the parasail entry sets it and
the generated Responses config overrides store=false on every call. This
keeps callers from hitting "State storage not supported" and matches what
Parasail's docs require.

Adds the PARASAIL enum value, listing under openai_compatible_providers,
provider documentation at docs/my-website/docs/providers/parasail.md, and
a focused unit test file under tests/test_litellm/llms/parasail/ that
covers JSON registration, chat URL construction, Responses URL
construction with PARASAIL_API_BASE override, and the force_store_false
regression in both the caller-sent-store=true and caller-omitted cases.

* fix(parasail): register in provider_endpoints_support, drop in-repo docs

Greptile review feedback. The provider doc belongs in the litellm-docs
repo, not this one's docs/my-website tree; removing it here. Adds the
parasail entry to provider_endpoints_support.json so the
check_provider_folders_documented.py CI check passes (chat_completions
and responses true; others false).

* fix: normalize Anthropic passthrough server tool usage (BerriAI#29827)

* test(anthropic): cover server_tool_use dict cost tracking

* fix: normalize Anthropic server tool usage

(cherry picked from commit 982f726)

* fix: keep server tool usage subscriptable

(cherry picked from commit 70280b9)

---------

Co-authored-by: Genmin <joey@joeyroth.com>

* fix(proxy): fix typo generic_role_mappoings -> generic_role_mappings in ui_sso.py (BerriAI#29753)

Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639) (BerriAI#29493)

* feat(proxy): add disable_budget_reservation general setting (BerriAI#27639)

* feat(proxy): register disable_budget_reservation in ConfigGeneralSettings (BerriAI#27639)

* docs(proxy): document disable_budget_reservation concurrency tradeoff (BerriAI#27639)

* ci: re-trigger flaky docker build (prisma generate ECONNRESET)

* fix(proxy): warn and document budget enforcement tradeoff when disable_budget_reservation is set (BerriAI#27639)

* feat(gemini_tts): adding support to Gemini TTS languageCode parameters (BerriAI#29623)

* Adding support to Gemini TTS Language Code parameters

* Mapping Gemini TTS languageCode param in Docstring

* Use snake_case for language_code input keyMapping Gemini TTS languageCode param in Docstring

* Restoring files modified under enterprise/litellm_enterprise due to lint/formatting checks

---------

Co-authored-by: João Garrido <joaogarrido@google.com>

* feat(guardrails): capture user and model metadata in CrowdStrike AIDR (BerriAI#29517)

* fix(proxy): require OpenAI path segment for shared Azure Cognitive Services domains

Address Greptile review: the `*.cognitiveservices.azure.com` /
`*.openai.azure.com` domains are shared by every Azure Cognitive Service
(Speech, Vision, Language, ...), so a hostname-only substring match
misclassified non-OpenAI Azure traffic as OpenAI routes.

- Replace the substring host test with suffix matching (rejects look-alike
  domains like cognitiveservices.azure.com.attacker.example).
- Add `_is_openai_compatible_url` that requires an OpenAI-style path marker
  (`/openai/` or `/v1/`) on the shared Azure domains, and use it in
  PassThroughEndpointLogging.is_openai_route (previously hostname-only).
- Add negative tests for Azure Speech/Vision paths and look-alike domains.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* fix: support Responses input in Redis semantic cache (BerriAI#29581)

* fix: support responses input in redis semantic cache

* test: cover redis semantic prompt extraction

* test: handle blank redis semantic text fallbacks

* chore: remove async cache dead statement

* test: cover redis semantic cache miss paths

* fix: filter sensitive cache lookup kwargs

* chore: rerun ci after huggingface rate limit

* chore(ui): regenerate dashboard API types (npm run gen:api)

Sync src/lib/http/schema.d.ts with the proxy OpenAPI spec: adds the
disable_budget_reservation general-settings field and picks up the
RateLimitError docstring reindent. Fixes the gen:api CI drift check.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* test(bedrock): assert empty additionalModelRequestFields is omitted

The Converse transformer now drops an empty additionalModelRequestFields
block instead of sending it as `{}`. Update test_bedrock_top_k_param so
models without top_k support (llama3) assert the key is absent rather than
equal to an empty dict.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Kent <72616338+kingdoooo@users.noreply.github.com>
Co-authored-by: codgician <15964984+codgician@users.noreply.github.com>
Co-authored-by: Praveen Ghuge <95286176+pghuge-cloudwiz@users.noreply.github.com>
Co-authored-by: Roi <roytev@gmail.com>
Co-authored-by: shin-berri <shin-laptop@berri.ai>
Co-authored-by: yuneng-jiang <yuneng@berri.ai>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Liam Scott <liam@uilliam.com>
Co-authored-by: abhay23-AI <abhaytrivedi22@gmail.com>
Co-authored-by: Ceder Dens <cederdens@gmail.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>
Co-authored-by: Kai Huang <kaihuang724@gmail.com>
Co-authored-by: rinto <54238243+ririnto@users.noreply.github.com>
Co-authored-by: Genmin <joey@joeyroth.com>
Co-authored-by: Arnav Bhilwariya <arnavbhilwariya0408@gmail.com>
Co-authored-by: Armaan Sandhu <74664101+Ar-maan05@users.noreply.github.com>
Co-authored-by: João Garrido <48538534+johngarrido@users.noreply.github.com>
Co-authored-by: João Garrido <joaogarrido@google.com>
Co-authored-by: Kenan Yildirim <kenan@kenany.me>
Co-authored-by: Dávid Balatoni <balcsida@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Bug]: Created cache key includes full request payload (messages/tools/response_format), causing huge Redis keys and possible timeouts

2 participants