What is Raven?
Raven is an open-source AI gateway that you deploy on your own infrastructure. It sits between your application and LLM providers, giving you a single API to manage costs, enforce guardrails, and monitor usage — with full control over your data.Self-Host with Docker
Deploy Raven on your infrastructure in under 5 minutes with Docker Compose.
Quickstart
Try the managed cloud version at raven.dev.
Why Self-Host?
Own Your Data
Your API keys, request logs, and usage data never leave your infrastructure.
No Vendor Lock-in
Fully open-source. Fork it, extend it, and run it however you want.
Zero Added Latency
Deploy in the same network as your application for the fastest possible routing.
Core Features
Unified API
One OpenAI-compatible endpoint for multiple LLM providers. Switch models without changing code.
Virtual Keys
Create scoped API keys with rate limits, budgets, and environment separation.
Routing Rules
Route requests to different models based on cost, latency, or custom conditions.
Guardrails
Block harmful content, detect PII, and enforce content policies.
Cost Control
Set budgets, track spend in real-time, and get alerts before you overspend.
Analytics
Monitor usage, costs, and performance across all providers and models.
For Developers
API Reference
Complete API documentation with examples for every endpoint.
OpenAI Compatibility
Drop-in replacement for the OpenAI SDK — change one line of code.