BerriAI/litellm v1.63.12-nightly

on GitHub

What's Changed

• Fixes bedrock modelId encoding for Inference Profiles by @omrishiv in #9123
• Aim Security post-call guardrails support by @hxtomer in #8356
• Litellm dev 03 12 2025 contributor prs p2 by @krrishdholakia in #9216
• Support bedrock Application inference profiles + Support guardrails on streaming responses by @krrishdholakia in #9274
• v1.63.11-stable release notes by @ishaan-jaff in #9279
• Infer aws region from bedrock application profile id by @krrishdholakia in #9281
• feat: make masterkey secret configurable by @mknet3 in #9288
• fix(utils.py): Prevents final chunk w/ usage from being ignored by @krrishdholakia in #9314
• Update prompt_caching.md to fix typo by @afspies in #9317
• fix(redis_cache.py): add 5s default timeout by @krrishdholakia in #9322
• Support reading litellm proxy response cost header in sdk + support setting lower ssl security level by @krrishdholakia in #9330
• [Bug fix] Reset Budget Job by @ishaan-jaff in #9329
• fix(ollama/completions/transformation.py): pass prompt, untemplated o… by @krrishdholakia in #9333
• [UI] - Allow controlling default internal user settings on ui by @ishaan-jaff in #9328
• [Patch] - Allow disabling all spend updates / writes to DB by @ishaan-jaff in #9331

New Contributors

• @afspies made their first contribution in #9317

Full Changelog: v1.63.11-nightly...v1.63.12-nightly

Docker Run LiteLLM Proxy

docker run \
-e STORE_MODEL_IN_DB=True \
-p 4000:4000 \
ghcr.io/berriai/litellm:main-v1.63.12-nightly

Don't want to maintain your internal proxy? get in touch 🎉

Hosted Proxy Alpha: https://calendly.com/d/4mp-gd3-k5k/litellm-1-1-onboarding-chat

Load Test LiteLLM Proxy Results