OpenAI Format

Veo Video Generation

Use `POST /v1/videos` to call the Veo model and submit asynchronous video tasks.

POSThttps://zx1.deepwl.net/v1/videos
Request
curl -X POST https://zx1.deepwl.net/v1/videos \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '
{
  "model": "veo_3_1-fast",
  "prompt": "Dance in the center of the square",
  "size": "720x1280",
  "seconds": "8",
  "input_reference": "data:image/png;base64,BASE64_IMAGE",
  "metadata": {
    "output_config": {
      "aspect_ratio": "9:16",
      "audio_generation": "Disabled",
      "resolution": "720P"
    }
  }
}'
Response
{
  "id": "video_abc123",
  "object": "video",
  "model": "veo_3_1-fast",
  "status": "queued",
  "progress": 0,
  "created_at": 1735689600,
  "seconds": "8",
  "size": "720x1280"
}

OpenAI-format entry point, with JSON submission as the primary method. If you need to use the unified video POST /v1/video/create, please see Create Video.

  • Currently, size will be converted and submitted together with metadata.output_config.resolution and aspect_ratio.
  • Reference images are still expressed by input_reference, but in JSON requests they are usually represented as a URL, base64, or Data URL.

Current Models

  • veo_3_1
  • veo_3_1-fast

Request Headers

Authorizationstring必填

Authentication header. Uses a Bearer token, e.g. Bearer YOUR_API_KEY.

Content-Typestring必填

Request content type, must be application/json.

Request Body

modelstring必填

Model name. Currently only veo_3_1 and veo_3_1-fast are organized.

promptstring必填

Prompt.

sizestring

Size string. Common values currently include 720x1280 or 1280x720, which are used to derive Veo's resolution and aspect ratio.

secondsstring | integer

Target duration. Currently submitted through seconds; the gateway also supports duration and metadata.durationSeconds.

input_referencestring | array<string>

Reference image input. JSON requests can currently use a URL, base64, or data URI; the Veo adapter layer also supports multipart file uploads.

metadataobject

Extended configuration. The Veo pipeline will continue parsing metadata into upstream Veo fields such as DurationSeconds, AspectRatio, and Resolution.

enable_upsampleboolean

This is the Apifox field reference you provided, but based on the current main repository pipeline, it is not an explicit top-level field in the generic Veo DTO. If you want to express clarity or resolution, size and metadata are currently recommended instead.

Request Examples

Response Examples

Current Rules

ItemRule
Request formatJSON
Reference image formatSupports URL, base64, and data URI, and also supports multipart files
Reference-to-video modeThe aspect ratio will be forcibly constrained to 16:9
Parameter mappingsize will be further mapped to Veo-required resolution and aspectRatio