What's Changed
- feat: show proxy url in ModelHub by @janfrederickk in #21660
- fix(bedrock): correct modelInput format for Converse API batch models by @hztBUAA in #21656
- fix: only tag selected deployment in access group creation by @hztBUAA in #21655
- feat(proxy): add custom favicon support by @hztBUAA in #21653
- fix(bedrock): prevent double UUID in create_file S3 key by @hztBUAA in #21650
- feat(semantic-cache): support configurable vector dimensions by @hztBUAA in #21649
- fix(utils): normalize camelCase thinking param keys to snake_case by @Chesars in #21762
- feat: add optional digest mode for Slack alert types by @dkindlund in #21683
- [Docs] store_model_in_db Release Docs by @yuneng-jiang in #21863
- Litellm dev 02 19 2026 p2 by @krrishdholakia in #21871
- fix(budget): fix timezone config lookup and replace hardcoded timezone map with ZoneInfo by @LeeJuOh in #21754
- fix: add missing return type annotations to iterator protocol methods in streaming_handler by @WhoisMonesh in #21750
- Add gollem Go agent framework cookbook example by @trevorprater in #21747
- fix: avoid mutating caller-owned dicts in SpendUpdateQueue aggregation by @themavik in #21742
- fix(vertex_ai): enable context-1m-2025-08-07 beta header by @edwiniac in #21870
- Revert "fix(vertex_ai): enable context-1m-2025-08-07 beta header" by @krrishdholakia in #21876
- fix: enable context-1m-2025-08-07 beta header for vertex_ai provider by @stakeswky in #21867
- Guardrail Policy Versioning by @krrishdholakia in #21862
- docs: add OpenClaw integration tutorial by @Chesars in #21605
- Litellm fix langfuse otel trace v2 by @Harshit28j in #21309
- fix(bedrock): encode model arns for OpenAI compatible bedrock imported models by @ta-stripe in #21701
- feat(bedrock): support optional regional STS endpoint in role assumption by @ta-stripe in #21640
- fix: ensure arrival_time is set before calculating queue time by @Harshit28j in #21918
- merge main in oss 22 02 by @Sameerlite in #21924
- merge main in oss 21 02 by @Sameerlite in #21926
- Add Noma guardrails v2 based on custom guardrails by @TomAlon in #21400
- Litellm dev 02 19 2026 p2 (#21871) by @krrishdholakia in #21872
- State management fixes for CheckBatchCost by @ephrimstanley in #21921
- Fix: Anthropic model wildcard access issue by @Sameerlite in #21917
- Litellm oss staging 02 22 2026 by @krrishdholakia in #21877
- Litellm oss staging 02 21 2026 by @krrishdholakia in #21786
- fix model cost map for anthropic fast and inference_geo by @Sameerlite in #21904
- Add Priority PayGo cost tracking gemini/vertex ai by @Sameerlite in #21909
- fix(tests): fix StopIteration in prisma self-heal cooldown test by @jquinter in #21938
- fix(tests): use absolute path for model_prices JSON validation test by @jquinter in #21939
- docs: add Google GenAI SDK tutorial (JS & Python) by @Chesars in #21885
- docs(ui): add pre-PR checklist to UI contributing guide by @Chesars in #21886
- feat: Litellm network mock by @ryan-crabbe in #21942
- Guardrail Monitor - measure guardrail reliability in prod by @krrishdholakia in #21944
New Contributors
- @janfrederickk made their first contribution in #21660
- @hztBUAA made their first contribution in #21656
- @LeeJuOh made their first contribution in #21754
- @WhoisMonesh made their first contribution in #21750
- @trevorprater made their first contribution in #21747
- @edwiniac made their first contribution in #21870
- @stakeswky made their first contribution in #21867
- @ta-stripe made their first contribution in #21701
Full Changelog: v1.81.14-nightly...litellm_presidio-dev-v1.81.15