Documentation

API reference

trilinkai exposes an OpenAI-compatible chat completions API. Use your API key from the console with any OpenAI SDK or plain HTTP.

Base URL and auth

Requests go to /v1 or the OpenRouter-compatible /api/v1 prefix on the same host as this console. Authenticate inference requests with a bearer token; the model catalogue is public.

Authorization: Bearer sk-trilinkai-...
Content-Type: application/json

Quickstart

curl https://your-gateway/api/v1/chat/completions \
  -H "Authorization: Bearer sk-trilinkai-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "fast-coding",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Streaming

Set stream: true to receive server-sent events.

curl -N https://your-gateway/api/v1/chat/completions \
  -H "Authorization: Bearer sk-trilinkai-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "fast-coding",
    "stream": true,
    "messages": [{"role": "user", "content": "Stream this"}]
  }'

Model catalogue

The public catalogue follows the OpenRouter model schema and does not require an API key.

curl https://your-gateway/api/v1/models

Models

Model	Best for
frontier-reasoning	Deep reasoning, planning, and hard debugging.
fast-coding	Low-latency coding and everyday chat.
long-context	Large documents and long conversations.
open-weights	Cheapest tier for high-volume jobs.

Errors

Status	Meaning
401	Missing, invalid, revoked, or blocked API key.
403	The key is not allowed to use this model.
429	Rate limit or budget exhausted.
5xx	Upstream provider error; retry with backoff.

Limits

20 requests/second per IP at the edge, burst 40.
120 requests/minute and 200k tokens/minute per API key.
Keys can be scoped to all models or a single tier.
Calls stop with 401 when the workspace balance reaches zero.

Create an API key Try the playground