mirror of
https://github.com/whit3rabbit/anyllm-proxy.git
synced 2026-09-21 16:00:49 +00:00
whit3rabbitandClaude Opus 4.8
65cf673b97
Large admin-ui refactor (Performative component system, sidebar nav,
provider/route tabs) plus backend module restructuring.
Admin UI contract fixes (this session):
- Fix Models page crash: useBackends unwraps {backends:[...]}; align
ModelEntry to {model_name, deployments} and ModelsResponse.strategy;
fix add-model body to {model_name, actual_model, backend_name}.
- Fix Backends/Providers health rendering: source per-backend status and
latency from the uptime endpoint (health_checks); narrow Backend type to
the real get_backends shape.
- Route + sidebar-link the previously-unrouted Backends tab.
Verified: cargo test (exit 0), clippy -D warnings (exit 0), fmt --check,
admin-ui tsc + vite build all green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
123 lines
8.8 KiB
Bash
123 lines
8.8 KiB
Bash
# anyllm-proxy configuration reference
|
|
# Copy to .env and fill in the values you need:
|
|
# cp .env.example .env
|
|
#
|
|
# Lines without a leading # are active defaults.
|
|
# Lines with # are optional or backend-specific — uncomment to use.
|
|
|
|
# ── Core / Server ─────────────────────────────────────────────────────────────
|
|
LISTEN_PORT=3000
|
|
# PROXY_CONFIG=/path/to/config.yaml # Simple YAML, LiteLLM YAML, or TOML
|
|
REQUEST_TIMEOUT_SECS=900 # 0 = disabled
|
|
# IP_ALLOWLIST=192.168.1.0/24,10.0.0.0/8 # Comma-separated CIDRs; bare IPs accepted
|
|
# TRUST_PROXY_HEADERS=true # Use X-Forwarded-For for client IP (requires IP_ALLOWLIST)
|
|
# WEBHOOK_URLS=https://example.com/hook # Comma-separated; fire-and-forget POST on request completion
|
|
|
|
# ── Auth / Access Control ─────────────────────────────────────────────────────
|
|
# Set exactly one of the following (or configure OIDC below).
|
|
# Without any, the proxy rejects all requests.
|
|
#
|
|
# PROXY_API_KEYS=sk-key1,sk-key2 # Comma-separated static keys
|
|
# LITELLM_MASTER_KEY=sk-... # LiteLLM alias for PROXY_API_KEYS
|
|
# PROXY_OPEN_RELAY=false # true = accept any non-empty key (DEV ONLY — insecure)
|
|
|
|
# ── Backend Selection ─────────────────────────────────────────────────────────
|
|
BACKEND=openai # openai | azure | vertex | gemini | anthropic | bedrock
|
|
|
|
# Model mapping — Anthropic model names are mapped to backend model names.
|
|
# Defaults depend on BACKEND (gpt-4o / gpt-4o-mini for OpenAI; gemini-2.5-pro / gemini-2.5-flash for Vertex/Gemini).
|
|
# BIG_MODEL=gpt-4o # Handles claude-sonnet and claude-opus requests
|
|
# SMALL_MODEL=gpt-4o-mini # Handles claude-haiku requests
|
|
|
|
# ── Backend: OpenAI ───────────────────────────────────────────────────────────
|
|
OPENAI_API_KEY=sk-...
|
|
# OPENAI_BASE_URL=https://api.openai.com # Override for local LLMs, vLLM, Ollama, etc.
|
|
# OPENAI_API_FORMAT=chat # chat (default, Chat Completions) | responses (Responses API)
|
|
|
|
# ── Backend: Azure OpenAI ─────────────────────────────────────────────────────
|
|
# BACKEND=azure
|
|
# AZURE_OPENAI_API_KEY=... # alias: AZURE_API_KEY
|
|
# AZURE_OPENAI_ENDPOINT=https://myresource.openai.azure.com # alias: AZURE_API_BASE
|
|
# AZURE_OPENAI_DEPLOYMENT=gpt4o
|
|
# AZURE_OPENAI_API_VERSION=2024-10-21 # alias: AZURE_API_VERSION
|
|
|
|
# ── Backend: Google Vertex AI ─────────────────────────────────────────────────
|
|
# BACKEND=vertex
|
|
# VERTEX_PROJECT=my-gcp-project
|
|
# VERTEX_REGION=us-central1
|
|
# VERTEX_API_KEY=... # or use GOOGLE_ACCESS_TOKEN (one required)
|
|
# GOOGLE_ACCESS_TOKEN=... # OAuth bearer token (alternative to VERTEX_API_KEY)
|
|
|
|
# ── Backend: Google Gemini ────────────────────────────────────────────────────
|
|
# BACKEND=gemini
|
|
# GEMINI_API_KEY=...
|
|
# GEMINI_BASE_URL=https://generativelanguage.googleapis.com/v1beta
|
|
|
|
# ── Backend: AWS Bedrock ──────────────────────────────────────────────────────
|
|
# BACKEND=bedrock
|
|
# AWS_REGION=us-east-1 # alias: AWS_REGION_NAME
|
|
# AWS_ACCESS_KEY_ID=...
|
|
# AWS_SECRET_ACCESS_KEY=...
|
|
# AWS_SESSION_TOKEN=... # Optional; for STS temporary credentials
|
|
|
|
# ── Backend: Anthropic Passthrough ────────────────────────────────────────────
|
|
# No translation — forwards Anthropic requests as-is to the upstream Anthropic API.
|
|
# BACKEND=anthropic
|
|
# ANTHROPIC_API_KEY=sk-ant-...
|
|
|
|
# ── TLS / mTLS ────────────────────────────────────────────────────────────────
|
|
# TLS_CLIENT_CERT_P12=/path/to/client.p12 # PKCS#12 client cert for mTLS to backend
|
|
# TLS_CLIENT_CERT_PASSWORD=... # Required if TLS_CLIENT_CERT_P12 is set
|
|
# TLS_CA_CERT=/path/to/ca.pem # PEM CA for verifying the backend server cert
|
|
|
|
# ── Admin UI ──────────────────────────────────────────────────────────────────
|
|
# Requires --webui or --admin CLI flag (or WEBUI=1 via Docker entrypoint).
|
|
# ADMIN_PORT=3001
|
|
# ADMIN_BIND=127.0.0.1 # Set to 0.0.0.0 in Docker (host port binding limits exposure)
|
|
# ADMIN_DB_PATH=admin.db # SQLite path; recommended: /data/admin.db in Docker
|
|
# ADMIN_TOKEN= # Set explicitly (min 32 chars recommended).
|
|
# # Generate: openssl rand -hex 32
|
|
# # If unset: auto-generated on each restart, written to ADMIN_TOKEN_PATH.
|
|
# # Docker detached mode: docker compose exec proxy cat /data/.admin_token
|
|
# ADMIN_TOKEN_PATH=.admin_token # Recommended: /data/.admin_token in Docker
|
|
# DISABLE_ADMIN=false # 1/true to force-disable even when --webui flag is passed
|
|
# WEBUI=1 # Docker entrypoint shorthand for --webui (not read by proxy directly)
|
|
|
|
# ── Logging / Debug ───────────────────────────────────────────────────────────
|
|
RUST_LOG=info # error | warn | info | debug | trace
|
|
# LOG_BODIES=false # Log full request/response bodies at debug level
|
|
# REDACT_SECRETS=false # Redact detected secrets before forwarding JSON/text requests
|
|
# ANYLLM_DEGRADATION_WARNINGS=false # Set x-anyllm-degradation header when features are silently dropped
|
|
|
|
# ── Rate Limiting (Redis) ─────────────────────────────────────────────────────
|
|
# Requires binary built with: --features redis
|
|
# REDIS_URL=redis://localhost:6379 # Docker compose: redis://redis:6379
|
|
# RATE_LIMIT_FAIL_POLICY=open # open = allow on Redis failure | closed/deny = reject 503
|
|
|
|
# ── OIDC / JWT Auth ───────────────────────────────────────────────────────────
|
|
# When set, Bearer tokens are validated against the OIDC discovery endpoint.
|
|
# OIDC_ISSUER_URL=https://accounts.google.com
|
|
# OIDC_AUDIENCE=anyllm-proxy # Defaults to issuer URL if unset (with a warning)
|
|
|
|
# ── Observability: Langfuse ───────────────────────────────────────────────────
|
|
# When both keys are set, all requests are traced to Langfuse automatically.
|
|
# LANGFUSE_PUBLIC_KEY=pk-lf-...
|
|
# LANGFUSE_SECRET_KEY=sk-lf-...
|
|
# LANGFUSE_HOST=https://cloud.langfuse.com # Override for self-hosted instances
|
|
|
|
# ── Observability: OpenTelemetry ──────────────────────────────────────────────
|
|
# Requires binary built with: --features otel
|
|
# OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4318
|
|
# OTEL_SERVICE_NAME=anyllm-proxy
|
|
# OTEL_TRACES_SAMPLER=parentbased_always_on
|
|
|
|
# ── Semantic Cache (Qdrant) ───────────────────────────────────────────────────
|
|
# Requires binary built with: --features qdrant
|
|
# QDRANT_URL=http://localhost:6333
|
|
# QDRANT_COLLECTION=anyllm_cache # Default collection name
|
|
|
|
# ── Model Pricing ─────────────────────────────────────────────────────────────
|
|
# MODEL_PRICING_FILE=/path/to/pricing.json
|
|
# JSON array of: {model_pattern, input_cost_per_token, output_cost_per_token, provider}
|
|
# Falls back to embedded pricing if the file is unreadable.
|