Vidu Video Generation · Overview

Document Version: v1.0.0 | Last Updated: 2026-06-11
This platform fully supports the official Vidu video generation APIs. Requests and responses are transparently proxied with identical parameter semantics.

Capabilities

Vidu video generation supports the following task types, each submitted through a dedicated endpoint:

Task Type	Endpoint	Description
Text-to-Video (T2V)	`POST /vidu/ent/v2/text2video`	Generate video from a text prompt
Image-to-Video (I2V)	`POST /vidu/ent/v2/img2video`	Generate video driven by a first-frame image
Reference-to-Video	`POST /vidu/ent/v2/reference2video`	Generate video with subject consistency from reference images/videos; supports subject libraries
Start-End Frame	`POST /vidu/ent/v2/start-end2video`	Generate a transition video from specified start and end frame images
Smart Multi-Frame	`POST /vidu/ent/v2/multiframe`	Generate long videos from multiple keyframe images
Scene Effect Template	`POST /vidu/ent/v2/template`	Generate effect videos based on preset scene templates
Template Story	`POST /vidu/ent/v2/template-story`	Generate a complete video from a story template in one click

Video generation is asynchronous: after submission, a task_id is returned immediately. Use the Query Tasks endpoint to poll the task status. Once the task succeeds, the response contains creations.url (video download URL, valid for 24 hours) — please download or transfer the file promptly.

Standard Workflow

Create Task (POST /vidu/ent/v2/{endpoint})
        │  Returns task_id + state: "created"
        ▼
Query Status (GET /vidu/ent/v2/tasks?task_ids={task_id})   ← Poll until success or failed
        │  Returns creations.url (valid for 24 hours)
        ▼
Download Video File

Endpoints

API	Method	URL
Text-to-Video	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/text2video`
Image-to-Video	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/img2video`
Reference-to-Video	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/reference2video`
Start-End Frame	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/start-end2video`
Smart Multi-Frame	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/multiframe`
Scene Effect Template	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/template`
Template Story	`POST`	`https://platform.dataeyes.ai/vidu/ent/v2/template-story`
Query Task List	`GET`	`https://platform.dataeyes.ai/vidu/ent/v2/tasks`

Path Rule: Append the channel prefix /vidu after the platform domain, then concatenate the original Vidu API path.
For example, official https://api.vidu.cn/ent/v2/text2video → platform https://platform.dataeyes.ai/vidu/ent/v2/text2video.

Authentication

All requests are authenticated via the HTTP Authorization header with a Bearer Token:

Authorization: Bearer {API_KEY}

{API_KEY} is the API key created in your platform console.

Note: The official Vidu API uses the Token prefix, but this platform uniformly uses the Bearer prefix. Do not send the Token prefix — it will be rejected with an "invalid token" error.

Request Headers

Header	Required	Description
`Authorization`	Yes	`Bearer {API_KEY}`
`Content-Type`	Yes	Fixed as `application/json`

Model Overview

Model	Text-to-Video	Image-to-Video	Reference	Start-End Frame	Smart Multi-Frame	Audio-Video Output	Duration Range	Default Resolution
viduq3-pro	✅	✅	✅	✅	—	✅ (enabled by default)	1–16s	720p
viduq3-turbo	✅	✅	✅	✅	—	✅ (enabled by default)	1–16s	720p
viduq3-pro-fast	—	✅	—	—	—	✅ (enabled by default)	1–16s	720p
viduq3-mix	—	—	✅ (non-subject only)	—	—	✅	3–16s	720p
viduq2-pro	—	✅	✅	✅	✅	Optional	1–10s	720p
viduq2-pro-fast	—	✅	—	✅	—	Optional	1–10s	720p
viduq2-turbo	—	✅	—	✅	✅	Optional	1–10s	720p
viduq2	✅	—	✅	—	—	—	1–10s	720p
viduq1	✅	✅	✅	✅	—	—	5s	1080p
viduq1-classic	—	✅	—	✅	—	—	5s	1080p
vidu2.0	—	✅	✅	✅	—	—	4s/8s	360p/720p

The supported models may vary by endpoint. Refer to each endpoint's documentation for the specific model list.

Task States

State	Description
`created`	Created successfully
`queueing`	Waiting in queue
`processing`	Processing
`success`	Generation succeeded
`failed`	Generation failed

Callback Notifications

All task creation endpoints support the callback_url parameter. When set, Vidu will send a POST callback request to the specified URL whenever the task state changes. The callback body structure is identical to the Query Tasks response. Callback states include processing, success, and failed, with up to 3 retries on failure.

Scene Effect Template

Template Story

Query Tasks