← Portfolio
Live in your browser

llm-gateway

A real FastAPI application, handling real requests — in this tab. Auth, per-key rate limiting, response caching, retries and cost accounting all execute exactly as they do in production. Hit the buttons and watch the middleware actually work.

Booting Python…
// Send requests

Poke the gateway

Each button is a real HTTP request through the real ASGI app. Watch the status codes — and send the same message twice to see the cache take over.

Caching
Auth
Resilience & limits
Read-only
live /metrics · 0 requests sent this session · rate limit set to 8 tokens, refilling 0.5/s so the 429 is reachable by hand