curl --request POST \
  --url https://api.aiid.edu.kg/ent/v2/img2video \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "viduq3-turbo",
  "prompt": "A cinematic product ad with smooth camera motion",
  "duration": 5,
  "resolution": "720p",
  "aspect_ratio": "16:9"
}
'

{
  "task_id": "<string>",
  "state": "created",
  "model": "<string>",
  "prompt": "<string>",
  "images": [
    "<string>"
  ],
  "duration": 123,
  "resolution": "<string>",
  "created_at": "<string>",
  "creations": "<string>"
}

Video generation Vidu

Create image-to-video (Vidu)

This series supports OpenAI video generation formats (see the link Video generation Sora compatible formats).

Vidu official video generation API; the public path follows the Vidu official specification.

viduq3-turbo, viduq2, and vidu2.0 support text-to-video, image-to-video, first-and-last-frame video generation, or reference-to-video according to their respective capabilities. They can also be accessed through the /v1/videos Sora compatible format.

Vidu Q4: vidu-q4 is fixed at 720p, and vidu-q4-540p is fixed at 540p. Supports generating videos from 3 to 16 seconds. At least 1 image is required, with up to 15 reference images supported. Reference generation can include up to 3 MP3 audio clips, each 3 to 12 seconds long and up to 50 MB. Pure text-to-video, specified first and last frames, and reference videos are not supported. Multiple images are used as joint references and do not indicate first or last frame positions. Tasks can be created via POST /v1/videos, POST /api/v3/contents/generations/tasks, or the Vidu formats POST /ent/v2/img2video and POST /ent/v2/reference2video. For the unified endpoint, mode=reference_images is recommended; use mode=reference_material when including audio. Reference audio can use audio_urls. The Seedance format can also use the audio_url field of the standard content; the audio Boolean field controls audio generation and is used separately from the reference audio URL list. The resolution must match the model name.

Provides the public request skeleton, task query structure, and recommended validation template for the Vidu official video generation format, while the public path retains the Vidu official endpoint format.

Supported models: viduq3-turbo, viduq2, vidu2.0, vidu-q4, vidu-q4-540p

Unified request fields:

model (string, required): Public model name for Vidu. Vidu Q4 supports only image-to-video and reference-to-video with images; Q3/Q2/V2 retain their respective capabilities.
prompt (string, optional): Video generation prompt; required for text-to-video, and passed as needed for image-to-video, start-and-end-frame, and reference-based video generation.
images (array[string], optional): Image input. Single-image image-to-video uses 1 image; Vidu Q4 reference generation uses 1–15 images, which are used as joint references. For older models, use first-and-last-frame images and reference image counts according to their respective capabilities.
videos (array[string], optional): Video subject input is available only according to the capabilities of older models; Vidu Q4 does not support reference videos.
subjects (array[object], optional): Subject-library format input for reference-based video generation; may include fields such as subject name, image, video, or voice timbre.
duration (integer, optional): Video duration, in seconds. viduq3-turbo commonly uses 5 seconds; set it according to the official range.
resolution (string, optional): Output resolution, common values are 540p, 720p, and 1080p.
aspect_ratio (string, optional): Output aspect ratio, commonly used for text-to-video/reference-based video generation, such as 16:9, 9:16, and 1:1.
seed (integer, optional): Random seed.
movement_amplitude (string, optional): Motion intensity, common values are auto, small, medium, and large.
audio (boolean, optional): Whether to enable direct audio-video output.
audio_urls (array[string], optional): Reference audio for Vidu Q4. Supports up to 3 MP3 files, each 3–12 seconds and no larger than 50 MB; separate from the audio Boolean switch.
off_peak (boolean, optional): Whether to use staggered generation.
watermark (boolean, optional): Whether to add a watermark.

Common model/mode differences:

Vidu 文生视频: Supports the viduq3-turbo, viduq2, and vidu2.0 models; official-format text-to-video.
Vidu 图生视频: Supports the viduq3-turbo, viduq2, and vidu2.0 models; official-format single-image image-to-video.
Vidu 首尾帧生视频: Supports the viduq3-turbo, viduq2, and vidu2.0 models; official-format first-and-last-frame video generation using two images.
Vidu 参考生视频: Supports the viduq3-turbo, viduq2, and vidu2.0 models; official-format multi-reference image-to-video.
Vidu Q4 720p 单图: Supports the vidu-q4 model; examples of one image or two images with two MP3 reference clips; multiple images are used as shared references, with no first or last frame specified. Replace the example asset URLs before calling.
Vidu Q4 540p 单图: Supports the vidu-q4-540p model; examples of one image or two images with two MP3 reference clips; multiple images are used as shared references, with no first or last frame specified. Replace the example asset URLs before calling.

POST

ent

img2video

curl --request POST \
  --url https://api.aiid.edu.kg/ent/v2/img2video \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "viduq3-turbo",
  "prompt": "A cinematic product ad with smooth camera motion",
  "duration": 5,
  "resolution": "720p",
  "aspect_ratio": "16:9"
}
'

{
  "task_id": "<string>",
  "state": "created",
  "model": "<string>",
  "prompt": "<string>",
  "images": [
    "<string>"
  ],
  "duration": 123,
  "resolution": "<string>",
  "created_at": "<string>",
  "creations": "<string>"
}

Authorizations

Authorization

string

header

required

Use Bearer Token authentication. Format: Authorization: Bearer sk-xxxxxx

Body

application/json

model

enum<string>

required

Public model name for Vidu. Vidu Q4 supports only image-to-video and reference-to-video with images; Q3/Q2/V2 retain their respective capabilities.

Available options:

viduq3-turbo,

viduq2,

vidu2.0,

vidu-q4,

vidu-q4-540p

Example:

"viduq3-turbo"

images

string<uri>[]

required

Image input. Single-image image-to-image generation produces 1 image; Vidu Q4 reference generation produces 1–15 images, with the images used as shared references. For legacy models, first/last frames and the number of references are used according to their respective capabilities.

Required array length: 1 element

Example:

["https://example.com/input.jpg"]

prompt

string

Video generation prompt; required for text-to-video, and optional for image-to-video, first-and-last-frame-to-video, and reference-to-video, depending on business needs.

Example:

"A cinematic product ad with smooth camera motion"

videos

string<uri>[]

Video subject input is available only for legacy model capabilities; Vidu Q4 does not support reference videos.

subjects

object[]

Enter in the same format as the video subject library reference generation input. It can include fields such as subject name, images, videos, or voice timbre.

Hide child attributes

subjects.name

string

subjects.images

string<uri>[]

subjects.videos

string<uri>[]

subjects.voice_id

string

subjects.server_id

string

auto_subjects

boolean

duration

integer

Video duration, in seconds. viduq3-turbo Commonly 5 seconds; set it within the official range.

Example:

5

resolution

enum<string>

Output resolution, common values 540p, 720p, and 1080p.

Available options:

540p,

720p,

1080p

Example:

"720p"

aspect_ratio

enum<string>

Output aspect ratio, commonly used for text-to-video/reference-to-video generation, such as 16:9, 9:16, and 1:1.

Available options:

16:9,

9:16,

4:3,

3:4,

1:1

Example:

"16:9"

seed

integer

Random seed.

movement_amplitude

enum<string>

Motion range, common values auto, small, medium, large.

Available options:

auto,

small,

medium,

large

Example:

"auto"

audio

boolean

Whether to enable direct audio/video output.

audio_urls

string<uri>[]

Vidu Q4 reference audio: up to 3 MP3 clips, each 3–12 seconds and no larger than 50 MB; separate from the audio boolean switch.

Maximum array length: 3

audio_type

enum<string>

Available options:

all,

speech_only,

sound_effect_only

voice_id

string

is_rec

boolean

bgm

boolean

payload

string

off_peak

boolean

Whether to use staggered generation.

watermark

boolean

Whether to add a watermark.

wm_position

enum<integer>

Available options:

1,

2,

3,

4

wm_url

string<uri>

callback_url

string<uri>

Response

200 - application/json

Task created successfully

task_id

string

required

The task ID returned when creating a task, used to query the task.

state

enum<string>

required

Task status. Common values: created, queueing, processing, success, failed.

Available options:

created,

queueing,

processing,

success,

failed

model

string

prompt

string

images

string[]

duration

integer

resolution

string

created_at

string

creations

string

Overview

Gemini Format for Image Generation

Image generationOpenAI DALL-E format

List models

Music generation task format

OpenAI Format (Responses)

OpenAI format (Chat)

Video generation Kling format

Video generation Omni and Veo formats

Video generation Seedance

Video generation Sora compatible formats

Video generation Vidu

Video generationHappyHorse and Wan

Create image-to-video (Vidu)

Authorizations

Body

Response