Quickstart - GENGEN API

This guide submits a text-to-video task and polls it until the generated video is ready.

Creating a generation task is billable. Confirm the selected model and controls before sending the request.

Create an API key

Open API Keys in the GENGEN Dashboard and create a key.Store the full key when it is shown. GENGEN does not display the secret again.

Submit a task

Replace the placeholder key and send the request:

curl --request POST \
  --url https://gengen.farm/api/gengen/v1/contents/generations/tasks \
  --header 'Authorization: Bearer gengen_live_xxxxxxxxxxxxxxxx' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "dreamina-seedance-2-0-260128",
    "mode": "text_to_video",
    "prompt": "A kitten yawns at the camera in warm morning light.",
    "controls": {
      "duration": 5,
      "resolution": "720p",
      "ratio": "16:9",
      "generateAudio": true,
      "watermark": false
    }
  }'

A successful submission returns a public task ID:

{
  "id": "seedance20:cgt-20260419-abcdef1234567890",
  "model": "dreamina-seedance-2-0-260128",
  "status": "pending",
  "outputs": {
    "videos": [],
    "images": []
  }
}

Retrieve the result

URL-encode the task ID when placing it in the path. For example, encode : as %3A.

curl --request GET \
  --url https://gengen.farm/api/gengen/v1/contents/generations/tasks/seedance20%3Acgt-20260419-abcdef1234567890 \
  --header 'Authorization: Bearer gengen_live_xxxxxxxxxxxxxxxx'

Continue polling while the task is pending, queued, running, or processing. When status is completed or succeeded, read the video URL from outputs.videos.

{
  "id": "seedance20:cgt-20260419-abcdef1234567890",
  "model": "dreamina-seedance-2-0-260128",
  "status": "completed",
  "outputs": {
    "videos": [
      "https://example.com/generated-video.mp4"
    ],
    "images": []
  },
  "createdAt": "2026-04-19T08:00:00.000Z",
  "updatedAt": "2026-04-19T08:01:30.000Z",
  "usage": {
    "completion_tokens": 100000,
    "total_tokens": 100000
  }
}

Seedance response details

Seedance 2.0 and 2.5 report successful tasks with status: "completed". Accept both completed and succeeded when handling video results across providers, and read generated media from the standard outputs.videos and outputs.images fields. When the upstream provider includes usage, GENGEN returns it as an optional compatibility field after the standard response fields. Its child keys retain the provider’s original names, including snake_case:

Field	Description
`usage.completion_tokens`	Provider-reported completion token count.
`usage.total_tokens`	Provider-reported total token count.

The token counts above are illustrative. The available usage fields depend on the upstream response; task creation and in-progress responses may omit usage. Requests with video references can still return only completion_tokens and total_tokens, so do not assume that usage includes separate input or video token counts. These counts describe provider usage, not the charged price. Other compatibility fields, such as task_id, video_url, and content.video_url, may also be present. Use the standard id and outputs fields in new integrations.

Provider output URLs can expire. Copy successful outputs to storage that you control.

Next steps

Configure an existing OpenAI SDK through the OpenAI-compatible LLM guide.
Learn how generation modes and media inputs work.
Review the create task endpoint.
Open the API Explorer for a guided workflow.

​Seedance response details

​Next steps

Seedance response details

Next steps