Developer API

Authentication

Organization service keys are sent as Authorization: Bearer. Keys are displayed once at creation, stored as hashes, and managed in the admin console with model permissions and monthly budgets. Keep keys on your server.

Chat completions

POST /v1/chat/completions uses the existing SEADS gateway with an allowed concrete model ID from GET /v1/models. Organization policies and usage metering apply. The response follows the OpenAI chat-completions shape; streaming and seads-auto are not available yet.

curl https://api.seads.io/v1/chat/completions \
  -H "Authorization: Bearer $SEADS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<model-id-from-/v1/models>",
    "messages": [{"role": "user", "content": "Hello"}],
    "stream": false
  }'

Streaming

Set stream to true to receive OpenAI-compatible server-sent completion chunks. The stream ends with data: [DONE].

Models and routing

GET /v1/models lists enabled chat models permitted for the key. Use a concrete ID to pin a model, or use SEADS One (seads-one) to select the lowest-cost eligible chat model.

Rate limits and errors

Each service key is limited to 60 requests per minute. Responses include X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers. Monthly budgets are enforced before a request reaches a provider.

401

Invalid or missing key

403

Model access denied

402

Organization payment overdue

429

Rate or budget limit reached

AI and data disclosure

Responses are generated by AI. SEADS orchestrates third-party models; it does not claim to create them. The developer API is intended to hide provider identity unless the organization enables transparency. A model and subprocessor disclosure covering API data is required before launch.