Create Video muse-video
Endpoint: POST https://4zapi.com/v1/videos
Text-to-Video / Image-to-Video (First Frame), outputs a 720P MP4 with sound. Asynchronous task: After submission, poll with 'Query Video muse-video', and download with 'Download Video muse-video' after completed.
| Parameter | Required | Description |
|---|---|---|
| model | Yes | Fixed muse-video |
| prompt | Yes | Scene description, in Chinese or English, up to 4000 characters |
| seconds | No | Integer (string or number) between 4–30, default is 4; billed based on this duration. The field name is seconds, not duration |
| size | No | 1280x720 (landscape) or 720x1280 (portrait), defaults to landscape if not filled |
| input_reference | No | First frame image (Image-to-Video): image URL or data URL, supports PNG/JPEG/WebP, does not support file ID; can also upload files directly using multipart/form-data |
| resolution | No | Must be 720p, can be omitted |
- Billing: Charged based on requested seconds, pre-deducted upon submission; automatic refund for task failure; returns 400 (code is
client_local_validation) for invalid parameters, no charge for unbuilt tasks. - Actual duration rounds up, will not be shorter than requested: 4,5→5 seconds; 6→6; 7,8→8; 9,10→10; 11–15→15; 16–20→20; 21–25→25; 26–30→30.
- For durations over 10 seconds, segments are generated and stitched together, changing scenes at the 10th and 20th seconds.
- Time consumption: approximately 1–2 minutes for under 10 seconds, about 3–7 minutes for 15–30 seconds; it is recommended to check every 10–20 seconds.
Parameters
Content-Type(header, string, required)
Request body (application/json)
model(string, required)prompt(string, required)seconds(string, required)size(string, required)
Request example
{
"model": "muse-video",
"prompt": "金色的枫叶在微风中飘落，阳光穿过树梢，电影感",
"seconds": "5",
"size": "1280x720"
}