Skip to content

Node Express with streaming

This example is a slightly more realistic service: your app exposes its own HTTP API and calls LunarGate behind the scenes.

Source code: github.com/lunargate-ai/gateway-examples/tree/main/node-express-moderate

Best for

Use this example when you want to see how LunarGate fits inside an app server, not just a one-off script.

What it demonstrates

  • POST /api/ask for normal JSON request/response flow
  • GET /api/stream?prompt=... for SSE streaming through the gateway
  • X-LunarGate-User and X-LunarGate-SessionId request headers
  • model="lunargate/auto" as the client-side model choice

Run it

cp .env.example .env
npm install
npm run dev

The app starts on http://127.0.0.1:3001.

Test the endpoints

JSON endpoint

curl -X POST http://127.0.0.1:3001/api/ask \
  -H "Content-Type: application/json" \
  -d '{"prompt":"Explain why a gateway helps in a multi-team AI platform."}'

Streaming endpoint

curl -N "http://127.0.0.1:3001/api/stream?prompt=Write%20a%20short%20haiku%20about%20LunarGate"

What to inspect

  • src or server files for the app-side streaming endpoint
  • .env.example for gateway base URL and model defaults
  • config-observability.yaml.example if you want request inspection in the LunarGate Dashboard on app.lunargate.ai

Why this example matters

Many teams do not call the gateway directly from a browser or CLI. They call it from another backend service.

This example shows the separation clearly:

  • your app keeps its own API contract
  • LunarGate stays behind that app boundary
  • streaming can still pass through cleanly