modelmux 0.5.0

ModelMux - high-performance Rust gateway that translates OpenAI-compatible API requests to Vertex AI (Claude), with streaming, tool calling, and production-grade reliability.
Documentation
SET_CORS_HEADERS=true
LOG_LEVEL=trace
HOST=localhost
PORT=3000

# Provider: use LLM_URL (overrides provider fields) OR Vertex-specific vars
# Option 1 — full URL override (no :rawPredict/:streamRawPredict suffix)
# LLM_URL="https://europe-west1-aiplatform.googleapis.com/v1/projects/MY_PROJECT/locations/europe-west1/publishers/anthropic/models/claude-sonnet-4@20250514"
# LLM_MODEL_DISPLAY_NAME="claude-sonnet-4"

# Option 2 — Vertex (Google-docs structure)
LLM_PROVIDER=vertex
VERTEX_REGION=europe-west1
VERTEX_PROJECT=my-gcp-project
VERTEX_LOCATION=europe-west1
VERTEX_PUBLISHER=anthropic
VERTEX_MODEL_ID=claude-sonnet-4@20250514
# LLM_MODEL_DISPLAY_NAME="claude-sonnet-4"

GCP_SERVICE_ACCOUNT_KEY="ewo...<base64 key>...=="
LLM_CONTEXT_SIZE=200000
LLM_WORD_LIMIT=150000
LLM_TEMPERATURE=0.9
LLM_MAX_TOKENS=8000