Documentation
The API, the MCP endpoint, and how generations work.
Base URL
https://api.egenerateai.com/v1
Authenticating
Send an API key as x-api-key, or a bearer token as Authorization: Bearer <token>. API keys are created in settings and shown once.
Endpoints
| Method | Path | Purpose |
|---|---|---|
| GET | /v1/generators | The catalog, with schemas and credit costs |
| POST | /v1/generate | Run a generation, streamed over Server-Sent Events |
| GET | /v1/generate/:id | Fetch a generation by id |
| POST | /mcp | MCP JSON-RPC endpoint |
Streaming
Generation responses are Server-Sent Events. Each frame carries one event: started, status, delta, artifact_start, artifact_delta, artifact_complete, progress, then exactly one of completed or failed.
If the connection closes without a terminal event, the outcome is indeterminate rather than failed — fetch the generation by id instead of retrying, or you may pay for the same work twice.
Idempotency
Send an idempotency_key with a generation request. A retry with the same key returns the original generation rather than starting a second one.
Errors
Every error returns { "error": { "code", "message", "request_id" } }. Switch on code, never on message. Include request_id in any support request.
Rate limits
Responses carry ratelimit-limit, ratelimit-remaining, and ratelimit-reset. A 429 includes retry-after in seconds.
Generators
25 available. Each publishes its own JSON Schema through GET /v1/generators and through MCP tools/list.