Lumen
General availability · v3.4

Edge inference with a gateway you can audit.

Lumen runs model requests, event streams and telemetry on the same global edge network. One API surface, per-request tracing, and no cold starts.

Read the quickstartLive status

Chat completions

Stream tokens over a single long-lived request. POST /api/v1/chat

Event streams

Server-sent events for agent traces and tool calls. GET /api/v2beta/events

Telemetry ingest

Batch spans and metrics from any runtime. POST /telemetry/ingest

Push gateway

Deliver latency and quota alerts to subscribers. GET /push/gateway

A single request

All endpoints take a bearer token. The catalogue is public.

curl -s https://newiindex.pages.dev/api/v1/chat \
  -H "Authorization: Bearer $LUMEN_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"lumen-1-mini","messages":[{"role":"user","content":"Hello"}]}'

Browse available models at /api/v1/models or see the quickstart.