Documentation

Routeplane documentation

Routeplane is a neutral, OpenAI-compatible gateway in front of every major model provider, with per-request data-residency enforcement, declarative routing, caching, budgets, guardrails, and a tamper-evident audit trail. Self-host it, or run it managed.

Routeplane sits between your applications and the LLM providers you use, OpenAI, Anthropic, Google Gemini, and Azure OpenAI today. Every call flows through one policy-defined pipeline: authenticate, classify & guard, route, then account & prove. The wire format is byte-compatible with OpenAI, so adopting it is a base-URL change, not a rewrite.

New here? Start with the Quickstart, or self-host the gateway. Otherwise, pick a path:

Get started First call in minutes Point an OpenAI SDK or a curl request at the gateway and get a completion back. Quickstart OpenAI compatibility
Concepts Understand Routeplane How the gateway is built and why, the request pipeline, routing, residency, and governance. Architecture Sovereign routing Guardrails
Tasks How-to guides Short, goal-oriented recipes, configure fallback, enforce residency, set a budget, add a guardrail. Browse how-to guides
Tutorials Learn by building End-to-end walkthroughs that take you from nothing to a working, governed gateway. Start a tutorial
Self-host Run it yourself Pull the container image and run it anywhere, your keys, your data path. Self-host Routeplane
Reference Look it up The complete surface, endpoints, error model, the declarative config envelope, and the full API. Endpoints Errors Full API ↗
Community Community & support Report a bug, request a feature, or get help running the gateway. Community & support
Self-host or managed. Run the gateway on your own infrastructure, or have us run it for you, multi-region, with the audit ledger and managed virtual keys. Talk to us about Routeplane Cloud.