Quickstart

Cursor fast path

Open Cursor Models, set the base URL to https://api.strixgate.dev/v1, paste your Strix API key, and add strix-auto. Add strix-quick as Fast mode for small, focused tasks. If the link does not open, use Cursor Settings → Models.

1. Set your key

Keep the key in an environment variable rather than source control.

export STRIX_API_KEY="your-key"

2. Make a request

curl https://api.strixgate.dev/v1/chat/completions \
  -H "Authorization: Bearer $STRIX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "strix-auto",
    "messages": [{"role": "user", "content": "Explain this failing test and suggest a fix."}],
    "max_tokens": 512
  }'

messages must be a non-empty array. A non-streaming response uses the familiar OpenAI chat-completion envelope. To force an intent lane, replace strix-auto with a model such as strix-coding, strix-deep, or strix-bug-hunt.

3. Verify the route

Completion responses expose routing metadata through CORS-readable headers including x-strix-lane, x-strix-reason, x-strix-tier, and x-strix-request-id. See routing evidence for the full practical set.