// event · library

LiteLLM v1.100.0: signed images, RAG ingest security, auto-router presets

library

vendor: BerriAIpublished: 2h agoconfidence: 0.85id: d8b04ddc-2fde-4b15-b648-212ed709d7c8

LiteLLM 1.100.0 ships cosign-signed Docker images, enforces vector-store upload security on /v1/rag/ingest, adds a Gemini Family auto-router preset, enables Entra ID / OAuth auth on all Azure AI Foundry routes, per-team New Relic trace routing, better Anthropic /v1/messages thinking-block round-tripping to OpenAI backends, Databricks cache-token billing, and many proxy/router/UI fixes.

capability delta

An agent can now route through a LiteLLM proxy with signed images, Entra ID auth on Azure AI Foundry, a Gemini auto-router preset, per-team New Relic tracing, and hardened /v1/rag/ingest vector-store uploads.

curator note

Big point release: supply-chain (cosign signing), RAG ingest security, Azure AI Foundry OAuth, auto-router presets, and many provider fidelity fixes relevant to agents proxying multi-model traffic.

// install

pip install litellm==1.100.0

pypi · litellm @ 1.100.0

// example

bash

# Verify the signed image, then run the proxy
cosign verify \
  --key https://raw.githubusercontent.com/BerriAI/litellm/v1.100.0/cosign.pub \
  ghcr.io/berriai/litellm:v1.100.0

docker run -p 4000:4000 \
  -e AZURE_AI_FOUNDRY_TENANT_ID=$TENANT \
  ghcr.io/berriai/litellm:v1.100.0 \
  --config /app/config.yaml

// links

litellmproxyrouterazure-ai-foundryragcosigngeminianthropic

source: github_release · https://github.com/BerriAI/litellm/releases/tag/v1.100.0