AllrouterOne gateway for every model you use
Bring your own provider keys, issue scoped API keys, and route every request through a single OpenAI-compatible endpoint — with usage, budgets and audit built in.
Everything that sits between your app and the model
Keys, routing, limits and observability — managed in one place, exposed through one familiar API.
Bring your own keys
Store provider credentials once, encrypted at rest. Your spend stays on your own provider accounts — Allrouter never resells tokens.
Scoped API keys
Issue per-project keys with their own model allow-lists and budgets. Rotate or revoke instantly without touching upstream credentials.
Usage & budgets
Track tokens, cost and latency per key, model and project. Set hard budget caps and get alerts before a runaway job drains an account.
Routing policies
Map friendly model names to real providers, set fallbacks, and shift traffic on price or availability — without redeploying your app.
Drop-in compatible
Point any OpenAI SDK at one base URL. Chat, embeddings and streaming work unchanged — switch providers by editing a string.
Audit & access
Every key, request and config change is logged. Projects and scoped keys keep access tightly controlled.
From your code to the right model in a single hop
One source, many destinations — exactly like our mark. Your app talks to Allrouter; Allrouter talks to every provider.
Point your SDK at one URL. Ship on every model.
Create a workspace, add a provider key, and make your first routed request in under five minutes.