## Create task
`POST /v1/video/generations`
```bash cURL theme={null}
curl --request POST \
--url https://www.anyfast.ai/v1/video/generations \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "wan3.0-video",
"input": {
"prompt": "A kitten runs across a moonlit rooftop while distant neon signs flicker, cinematic lighting and a smooth tracking shot."
},
"parameters": {
"resolution": "1080P",
"ratio": "16:9",
"duration": 8,
"audio": true,
"seed": 42,
"watermark": false
}
}'
```
## Request headers
Bearer authentication in the form `Bearer YOUR_API_KEY`.
Request body format.
## Request body
AnyFast model ID. Use `wan3.0-video`, or use the AnyFast-specific `wan3.0-video-nsfw` variant in the `Direct` resource group. Alibaba's official page lists only the standard model.
Basic generation input. At least one of `input.prompt` or `input.media` must be present.
Video description in Chinese or English. Content beyond 20,000 characters is automatically truncated.
Video generation controls.
Output resolution: `1080P`, `720P`, or `480P`.
Output aspect ratio: `adaptive`, `16:9`, `4:3`, `1:1`, `3:4`, or `9:16`.
Output duration in seconds. With no video input, use `-1` for smart duration or an integer from 2 through 30.
Whether the generated video contains an audio track. The price is the same for `true` and `false`.
Random seed from `0` through `2147483647`.
Whether to add a watermark to the generated video.
## Create response
```json 200 theme={null}
{
"id": "asyntask_wan30example123",
"task_id": "asyntask_wan30example123",
"object": "video",
"model": "wan3.0-video",
"status": "queued",
"progress": 0,
"created_at": 1787280000
}
```
Task ID used to query the result.
Task ID, identical to `id`.
Returned object type, `video`.
AnyFast model ID used by the task.
Initial task status, normally `queued`.
Initial task progress percentage.
Creation time as a Unix timestamp in seconds.
## Create task
`POST /v1/video/generations`
```bash cURL theme={null}
curl --request POST \
--url https://www.anyfast.ai/v1/video/generations \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "wan3.0-video",
"input": {
"prompt": "The painted character steps away from the wall and begins an energetic rap performance as the camera moves closer.",
"media": [
{
"type": "first_frame",
"url": "https://example.com/first-frame.png"
}
]
},
"parameters": {
"resolution": "1080P",
"ratio": "adaptive",
"duration": 5,
"audio": true,
"seed": 42,
"watermark": false
}
}'
```
## Request headers
Bearer authentication in the form `Bearer YOUR_API_KEY`.
Request body format.
## Request body
Use `wan3.0-video`, or the AnyFast-specific `wan3.0-video-nsfw` variant in the `Direct` resource group.
Basic generation input. At least one of `input.prompt` or `input.media` must be present.
Motion, scene, camera, and audio description. Content beyond 20,000 characters is automatically truncated.
Supply exactly one `first_frame` image. Do not combine frame media with `reference_image`, `reference_video`, `reference_audio`, `file`, or `link`.
Use `first_frame`.
Public image URL or Base64 data URI.
Video generation controls.
Output resolution: `1080P`, `720P`, or `480P`.
Output aspect ratio: `adaptive`, `16:9`, `4:3`, `1:1`, `3:4`, or `9:16`. `adaptive` can follow the source image and intent.
Output duration in seconds. Use `-1` for smart duration or an integer from 2 through 30.
Whether the generated video contains an audio track. The price is the same for `true` and `false`.
Random seed from `0` through `2147483647`.
Whether to add a watermark to the generated video.
## Create response
```json 200 theme={null}
{
"id": "asyntask_wan30example123",
"task_id": "asyntask_wan30example123",
"object": "video",
"model": "wan3.0-video",
"status": "queued",
"progress": 0,
"created_at": 1787280000
}
```
Task ID used to query the result.
Task ID, identical to `id`.
Returned object type, `video`.
AnyFast model ID used by the task.
Initial task status, normally `queued`.
Initial task progress percentage.
Creation time as a Unix timestamp in seconds.
## Create task
`POST /v1/video/generations`
```bash cURL theme={null}
curl --request POST \
--url https://www.anyfast.ai/v1/video/generations \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "wan3.0-video",
"input": {
"prompt": "The girl moves from a quiet smile to laughter while the camera slowly pushes in and the light changes from cool to warm.",
"media": [
{
"type": "first_frame",
"url": "https://example.com/first-frame.jpg"
},
{
"type": "last_frame",
"url": "https://example.com/last-frame.jpg"
}
]
},
"parameters": {
"resolution": "1080P",
"ratio": "adaptive",
"duration": 5,
"audio": true,
"seed": 42,
"watermark": false
}
}'
```
## Request headers
Bearer authentication in the form `Bearer YOUR_API_KEY`.
Request body format.
## Request body
Use `wan3.0-video`, or the AnyFast-specific `wan3.0-video-nsfw` variant in the `Direct` resource group.
Basic generation input. At least one of `input.prompt` or `input.media` must be present.
Motion and transition description. Content beyond 20,000 characters is automatically truncated.
Supply one `first_frame` and one `last_frame`. Do not combine them with `reference_image`, `reference_video`, `reference_audio`, `file`, or `link`.
Use `first_frame` for the opening image and `last_frame` for the ending image. Each type allows at most one item.
Public image URL or Base64 data URI.
Video generation controls.
Output resolution: `1080P`, `720P`, or `480P`.
Output aspect ratio: `adaptive`, `16:9`, `4:3`, `1:1`, `3:4`, or `9:16`. `adaptive` can follow the frame inputs and intent.
Output duration in seconds. Use `-1` for smart duration or an integer from 2 through 30.
Whether the generated video contains an audio track. The price is the same for `true` and `false`.
Random seed from `0` through `2147483647`.
Whether to add a watermark to the generated video.
## Create response
```json 200 theme={null}
{
"id": "asyntask_wan30example123",
"task_id": "asyntask_wan30example123",
"object": "video",
"model": "wan3.0-video",
"status": "queued",
"progress": 0,
"created_at": 1787280000
}
```
Task ID used to query the result.
Task ID, identical to `id`.
Returned object type, `video`.
AnyFast model ID used by the task.
Initial task status, normally `queued`.
Initial task progress percentage.
Creation time as a Unix timestamp in seconds.
## Create task
`POST /v1/video/generations`
```bash cURL theme={null}
curl --request POST \
--url https://www.anyfast.ai/v1/video/generations \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "wan3.0-video",
"input": {
"prompt": "Image 1 performs the motion from Video 1 while the scene follows the rhythm of Audio 1.",
"media": [
{
"type": "reference_image",
"url": "https://example.com/character.jpg"
},
{
"type": "reference_video",
"url": "https://example.com/motion.mp4"
},
{
"type": "reference_audio",
"url": "https://example.com/rhythm.mp3"
}
]
},
"parameters": {
"resolution": "1080P",
"ratio": "16:9",
"duration": 10,
"audio": true,
"seed": 42,
"watermark": false
}
}'
```
## Request headers
Bearer authentication in the form `Bearer YOUR_API_KEY`.
Request body format.
## Request body
Use `wan3.0-video`, or the AnyFast-specific `wan3.0-video-nsfw` variant in the `Direct` resource group.
Basic generation input. At least one of `input.prompt` or `input.media` must be present.
Video instruction. Content beyond 20,000 characters is automatically truncated. Refer to assets by their type-specific order, such as `Image 1`, `Video 1`, and `Audio 1`.
Reference assets. Supply up to 10 images, 5 videos, and 5 audio clips. Do not combine reference media with `first_frame` or `last_frame`.
Use `reference_image`, `reference_video`, or `reference_audio`.
Public media URL. Images also accept Base64 data URIs.
Video generation controls.
Output resolution: `1080P`, `720P`, or `480P`.
Output aspect ratio: `adaptive`, `16:9`, `4:3`, `1:1`, `3:4`, or `9:16`.
Output duration in seconds. Use `-1` for smart duration. Without a video input, use 2–30; with video input, total input-video duration plus output duration must not exceed 30 seconds.
Whether the generated video contains an audio track. The price is the same for `true` and `false`.
Random seed from `0` through `2147483647`.
Whether to add a watermark to the generated video.
## Create response
```json 200 theme={null}
{
"id": "asyntask_wan30example123",
"task_id": "asyntask_wan30example123",
"object": "video",
"model": "wan3.0-video",
"status": "queued",
"progress": 0,
"created_at": 1787280000
}
```
Task ID used to query the result.
Task ID, identical to `id`.
Returned object type, `video`.
AnyFast model ID used by the task.
Initial task status, normally `queued`.
Initial task progress percentage.
Creation time as a Unix timestamp in seconds.
## Create task
`POST /v1/video/generations`
```bash File theme={null}
curl --request POST \
--url https://www.anyfast.ai/v1/video/generations \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "wan3.0-video",
"input": {
"prompt": "Create a concise product-launch video from the supplied presentation, emphasizing the design details and three core benefits.",
"media": [
{
"type": "file",
"url": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260806/ebapmr/glass.pptx"
}
]
},
"parameters": {
"resolution": "1080P",
"ratio": "16:9",
"duration": 10,
"audio": true,
"seed": 42,
"watermark": false
}
}'
```
```bash Webpage theme={null}
curl --request POST \
--url https://www.anyfast.ai/v1/video/generations \
--header 'Authorization: Bearer YOUR_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
"model": "wan3.0-video",
"input": {
"prompt": "Turn the main points on this public page into a concise narrated video.",
"media": [
{
"type": "link",
"url": "https://example.com/public-product-overview"
}
]
},
"parameters": {
"resolution": "1080P",
"ratio": "16:9",
"duration": 10,
"audio": true,
"seed": 42,
"watermark": false
}
}'
```
## Request headers
Bearer authentication in the form `Bearer YOUR_API_KEY`.
Request body format.
## Request body
Use `wan3.0-video`, or the AnyFast-specific `wan3.0-video-nsfw` variant in the `Direct` resource group.
Basic generation input. At least one of `input.prompt` or `input.media` must be present.
Video instruction. Content beyond 20,000 characters is automatically truncated.
Supply at most one file or one webpage link. `file` and `link` are mutually exclusive, and neither can be combined with `first_frame` or `last_frame`.
Use `file` for a document or `link` for a public webpage.
Public HTTP or HTTPS URL. Webpages must be accessible without login.
Video generation controls.
Output resolution: `1080P`, `720P`, or `480P`.
Output aspect ratio: `adaptive`, `16:9`, `4:3`, `1:1`, `3:4`, or `9:16`.
Output duration in seconds. Use `-1` for smart duration or an integer from 2 through 30.
Whether the generated video contains an audio track. The price is the same for `true` and `false`.
Random seed from `0` through `2147483647`.
Whether to add a watermark to the generated video.
## Create response
```json 200 theme={null}
{
"id": "asyntask_wan30example123",
"task_id": "asyntask_wan30example123",
"object": "video",
"model": "wan3.0-video",
"status": "queued",
"progress": 0,
"created_at": 1787280000
}
```
Task ID used to query the result.
Task ID, identical to `id`.
Returned object type, `video`.
AnyFast model ID used by the task.
Initial task status, normally `queued`.
Initial task progress percentage.
Creation time as a Unix timestamp in seconds.
## Query task by ID
`GET /v1/video/generations/{id}`
```bash cURL theme={null}
curl --request GET \
--url https://www.anyfast.ai/v1/video/generations/asyntask_wan30example123 \
--header 'Authorization: Bearer YOUR_API_KEY'
```
## Request headers
Bearer authentication in the form `Bearer YOUR_API_KEY`.
## Path parameters
Task ID returned by a create request.
Alibaba Cloud reports a typical generation time of 1–5 minutes and recommends polling at roughly 15-second intervals. This upstream estimate is not an AnyFast service-level commitment; actual processing time can vary. Tasks can be queried for 24 hours after creation, and successful result URLs remain valid for 24 hours.
## Query response
```json 200 theme={null}
{
"code": "success",
"message": "",
"data": {
"task_id": "asyntask_wan30example123",
"action": "generate",
"status": "SUCCESS",
"fail_reason": "https://example.com/generated-video.mp4",
"result_url": "https://example.com/generated-video.mp4",
"submit_time": 1787280000,
"start_time": 1787280005,
"finish_time": 1787280125,
"progress": "100%",
"request_id": "request_wan30example123",
"data": {
"output": {
"end_time": "2026-08-21 08:02:05.000",
"orig_prompt": "A kitten runs across a moonlit rooftop while distant neon signs flicker.",
"scheduled_time": "2026-08-21 08:00:05.000",
"submit_time": "2026-08-21 08:00:00.000",
"task_id": "upstream_task_wan30example123",
"task_status": "SUCCEEDED",
"video_url": "https://example.com/generated-video.mp4"
},
"request_id": "upstream_request_wan30example123",
"usage": {
"SR": 1080,
"duration": 8,
"fps": 30,
"input_video_duration": 0.0,
"output_video_duration": 8.0,
"ratio": "16:9",
"video_count": 1
}
}
}
}
```
Response code. `success` means the query was accepted.
Response message. A successful query normally returns an empty string.
Task status and result payload.
AnyFast task ID.
Task action. Wan3.0 creation tasks return `generate`.
Task status: `NOT_START`, `QUEUED`, `IN_PROGRESS`, `SUCCESS`, or `FAILURE`. Use this field to determine whether the task failed.
Compatibility field. Some successful responses repeat `result_url` here, as shown above; a non-empty value does not mean the task failed. Use `status` to determine the outcome and read the result from `result_url`. On `FAILURE`, this field can contain failure details.
Temporary generated-video URL returned when `status` is `SUCCESS`. This field can be omitted before the task completes.
Task submission time as a Unix timestamp in seconds.
Task start time as a Unix timestamp in seconds; `0` while the task is still queued.
Task completion time as a Unix timestamp in seconds; `0` until the task reaches a terminal state.
Task progress as a percentage string, such as `100%`.
AnyFast request trace ID.
Raw upstream response data.
Raw upstream task output.
Upstream task completion time. Present after successful completion.
Prompt received by the upstream service. Present after successful completion.
Upstream task scheduling time. Appears once processing begins and can be omitted while the task is queued.
Upstream task submission time.
Upstream task ID.
Upstream task status.
Generated-video URL in the raw upstream output. Present after successful completion.
Upstream request trace ID.
Upstream video usage details. Present after successful completion and omitted while the task is queued or processing.
Upstream output resolution value.
Generated duration in seconds.
Generated video frame rate.
Total input-video duration in seconds.
Output-video duration in seconds.
Generated video aspect ratio.
Number of generated videos.
## Media limits
### Images
* Roles: `first_frame`, `last_frame`, and `reference_image`.
* Formats: JPEG, JPG, PNG without an alpha channel, BMP, or WEBP.
* Each side must be 240–8,000 px, the aspect ratio must not exceed 8:1, and each image must be no larger than 20 MB.
* Images accept a public HTTP or HTTPS URL or a Base64 data URI. Supply at most one first frame, one last frame, or 10 reference images.
### Videos
* Role: `reference_video`; up to five clips with no more than 15 seconds total.
* Formats: MP4 or MOV. Each clip must be 1–15 seconds, each side must be 240–4,096 px, the aspect ratio must not exceed 8:1, and each file must be no larger than 100 MB.
* Videos require a public HTTP or HTTPS URL.
### Audio
* Role: `reference_audio`; up to five clips with no more than 15 seconds total.
* Formats: WAV or MP3. Each clip must be 1–15 seconds and no larger than 15 MB.
* Audio requires a public HTTP or HTTPS URL.
### Files and webpages
* `file` supports DOCX, DOC, XLSX, XLS, PPTX, PPT, PDF, TXT, KEY, PAGES, NUMBERS, and MD files up to 100 MB.
* PDF, DOCX, DOC, PPTX, PPT, KEY, and PAGES inputs support up to 50 pages.
* `link` accepts a public HTTP or HTTPS webpage that does not require login.
* Supply at most one `file` or one `link`; they are mutually exclusive.