02-Create-Video-Regeneration
Create Video Regeneration Task
Version: v1.0.0 | Last Updated: 2026-08-13 This platform fully supports the MiniMax Video Generation V2 (MiniMax-H3) official API. Requests and responses are transparently proxied, and all parameter semantics remain consistent with the official API.
task_id. The task's task_type is regeneration, and it is managed through the shared H3 Query Task, List Tasks, and Cancel or Delete Task endpoints.POST https://platform.dataeyes.ai/hailuo/v2/video_regenerationNote: This endpoint only regenerates videos that conform to the MiniMax-H3 768P output specification into 2K. It is not a general-purpose upscaler for arbitrary videos.
source_task_id and content (with base_video) are mutually exclusive and exactly one must be provided (providing both or neither returns a parameter error). The request body has no duration field — the output duration is inherited from the source video.Mode 1: Regenerate by Task ID (source_task_id)
task_id of a previous successful /v2/video_generation task to regenerate from its output. No content is needed.| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
model | string (enum) | Yes | — | Model name. Currently: MiniMax-H3 |
source_task_id | string | Yes | — | Source task ID. Constraints: requires allowlist access (contact the platform to enable); the source task must belong to your account, have status=succeeded, and have been created within the last 7 days |
resolution | string (enum) | Yes | — | Target resolution; currently fixed at 2K |
callback_url | string | No | — | Webhook URL for status changes; behaves the same as in Create Video Generation Task |
aigc_watermark | boolean | No | false | Whether to add an AIGC watermark |
Mode 2: Regenerate by Source Video (base_video)
type=video_url and role=base_video in content, together with all the original inputs used to generate that 768P video.| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
model | string (enum) | Yes | — | Model name. Currently: MiniMax-H3 |
content | array | Yes | — | Regeneration input array (see below) |
resolution | string (enum) | Yes | — | Target resolution; currently fixed at 2K |
callback_url | string | No | — | Same as above |
aigc_watermark | boolean | No | false | Same as above |
content array must include:text must be the final prompt actually sent to the model at that time (if H3-Context-IR was used, the enhanced prompt — not the original pre-enhancement prompt; up to 40,000 characters per text). All reference images, videos, and audio must also be identical to those used at generation time. Any mismatch may prevent the regeneration from achieving the expected result.type=video_url with role=base_video — exactly one is required.base_video Source Video Specification
base_video must conform to the MiniMax-H3 768P output specification:| Item | Specification |
|---|---|
| Audio track | Must contain an audio track; videos without audio are not supported |
| Frame rate | 24 fps |
| Width / height | Both must be divisible by 32 |
| Area (width × height) | ≤ 768 × 1344 (1,032,192 pixels) |
| Total frames | 107-362 frames, in increments of 17 frames (approx. 4-15 seconds) |
Request Examples
Response Example
{
"task_id": "424010985738631"
}| Field | Type | Description |
|---|---|---|
task_id | string | Regeneration task ID; use it with Query Task to retrieve status and results |
Billing
usage.total_seconds) at the regeneration unit price; the first 5 input images per task are free, additional images are billed per image. A 5-second estimate is pre-charged at submission and settled against actual duration on completion (difference charged or refunded); failed tasks are fully refunded. See Overview · Billing & Settlement.