Tạo tác vụ khung đầu-cuối thành video
- Official docs: https://platform.vidu.cn/docs/start-end-frame-to-video
- Create start/end frame to video task; returns task_id or details
- Control generation with request parameters
Authorizations
Request Body
Text prompt
Text description for video generation.
Note: Max 2000 characters
Model name. Allowed: viduq2, viduq1, viduq3-turbo, viduq3-pro
- viduq3-turbo: Faster than viduq3-pro
- viduq3-pro: High-quality audio-video
- viduq2: Latest model
- viduq1: Clear visuals, smooth transitions
Duration by model:
- viduq3-pro, viduq3-turbo: default 5, range 1–16
- viduq2: default 5s, 1–10
- viduq1: default 5s, 5 only
Random seed
If omitted or 0, random; else uses set value
Motion amplitude (see resolution field in source API for model-specific options)
viduq2 (1–10s): default 720p; 540p, 720p, 1080p
viduq1 (5s): 1080p
Passthrough parameter
Not processed; data transfer only
Note: Max 1048576 characters
Add watermark: true / false
Note 1: Fixed AI watermark; off by default
Note 2: Use watermarked_url from query API
Watermark position: 1–4 corners. Default 3
Watermark image URL
Metadata JSON passthrough; Vidu default if empty
Callback URL (POST). Status: processing, success, failed. Signature auth
Images: first = start frame, second = end frame
Note 1: Similar resolution ratio 0.8–1.25; aspect under 1:4 or 4:1
Note 2–5: Base64/URL, formats, 50 MB, POST 20 MB
data:image/png;base64,{base64_encode}
Use recommended prompt
- true: system recommends prompt (+10 credits)
- false: use input prompt
Resolution by model and duration (see Vidu docs for full matrix)
Background music, default false; true picks from BGM library
Responses
Thành công