API reference

All generation endpoints are asynchronous: they return a task ID, and you retrieve the result separately.

Base URL and auth#

curl -X POST https://api.dev.runwayml.com/v1/image_to_video \
  -H "Authorization: Bearer $RUNWAYML_API_SECRET" \
  -H "Content-Type: application/json" \
  -d '{ "model": "gen4.5", "promptText": "A sunrise", "ratio": "1280:720", "duration": 5 }'

The request lifecycle#

1. POST to a generation endpoint — returns { "id": "..." } immediately.
2. GET /v1/tasks/{id} — poll until SUCCEEDED, FAILED or CANCELED.
3. Read output[0] and download it within 24–48 hours before the URL expires.

The SDKs collapse steps 1 and 2 into .waitForTaskOutput(). See waiting for output.

Related#

• Making API calls — working code for every endpoint
• Models — valid model values
• HTTP errors — status codes and retry safety