Wan video generation

The wan3.0-video series submits async video tasks with a JSON body. Generation parameters go in the object.

POST /v1/videos

• Supports text-to-video, first-frame video, first-and-last-frame video, reference image/video/audio, reference file, and reference webpage modes.
• First-and-last-frame mode is mutually exclusive with reference-asset mode; and cannot be used together.
• Async processing: submission returns id or task_id.

Available models

Headers

Name	Type	Required	Description
	string	Yes	Request authentication. Use a Bearer Token, for example:
	string	Yes	Request content type. Must be set to

Request body

Name	Type	Required	Description
	string	Yes	Model name
	string	No	Video prompt (Chinese or English). Provide at least one of or
	object	Yes	Media assets and video generation parameters

metadata

Name	Type	Required	Description
	array<object>	No	Media asset list. Each item has and (public URL, temporary OSS URL, or image Base64 Data URL). Provide at least one of this field or top-level
	string	No	Resolution: , , . Default . Lowercase is auto-uppercased
	string	No	Aspect ratio: , , , , , . Default
	integer	No	Output duration: without a reference video, range –, default . Pass for smart duration
	boolean	No	Whether to generate an audio track. Default
	integer	No	Random seed, range –
	boolean	No	Smart prompt rewrite. Default
	boolean	No	Whether to add a watermark. Default

Media asset types

	Max count	Description
	1	First-frame image, used strictly as the first video frame
	1	Last-frame image, used strictly as the last video frame
	10	Reference image
	5	Reference video; total duration ≤ 15 seconds
	5	Reference audio; total duration ≤ 15 seconds
	1	Reference file; cannot be used with
	1	Public webpage; cannot be used with

Asset limits: images (JPEG/JPG/PNG/BMP/WEBP; PNG transparency not supported; ≤20 MB; each side 240–8000 px; aspect ratio ≤ 8:1); video (MP4/MOV; ≤100 MB; each clip 1–15 s); audio (WAV/MP3; ≤15 MB; each clip 1–15 s); files (DOCX/XLSX/PPTX/PDF/TXT/MD, etc.; ≤100 MB; paginated formats ≤ 50 pages); webpages (public HTTP/HTTPS that do not require login).

Request example

Text-to-video

curl -X POST https://octopusx.ai/v1/videos \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "wan3.0-video",
  "prompt": "A kitten runs across a moonlit rooftop with city neon lights in the distance, cinematic quality, smooth camera motion.",
  "metadata": {
    "resolution": "1080P",
    "ratio": "16:9",
    "duration": 5,
    "audio": true,
    "seed": 123456789,
    "prompt_extend": true,
    "watermark": false
  }
}'

Response example

{
  "id": "task_1234567890abcdef",
  "object": "video",
  "model": "wan3.0-video-prime",
  "status": "queued",
  "progress": 0,
  "created_at": 1787673600,
  "metadata": {
    "url": ""
  }
}

Response fields

Name	Type	Description
	string	Task ID for later polling
	string	Always
	string	Model actually used
	string	Task status: , , , etc.
	integer	Progress percentage
	integer	Creation timestamp
	string	Result video URL when complete

Related endpoints

• Video APIs overview
• Wan task query
• Error codes