万相2.6 (Wanx 2.6) Reference-to-Video API

Generate character-consistent video from multimodal references (image, video). Async task API: submit a request, receive a task_id, then poll for status to get the result.

Supports multimodal input (image, video). Use a person or object as the main character to generate single-character performance or multi-character interaction videos. Up to 10 seconds and 1080P resolution.

This is an async task-based API: submit a request to receive a task_id, poll for status, then download the video when complete.

Base URL

https://api.autorouter.top — Production

Authentication

BearerAuth: http (bearer) Authenticate using a Bearer Token. Format: Authorization: Bearer sk-xxxxxx

Endpoints

POST `/api/v1/services/aigc/video-generation/video-synthesis`

Create Video Generation Task

Create a video generation task from reference images and reference videos.

The request header must include X-DashScope-Async: enable.

Request Body

Content-Type: application/json

Example:

{
  "model": "wan2.6-r2v-flash",
  "input": {
    "prompt": "Character2 sits on a chair by the window, holding character3, and plays a gentle American country ballad next to character4. Character1 says to Character2: \"That sounds nice\"",
    "reference_urls": [
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/hfugmr/wan-r2v-role1.mp4",
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/qigswt/wan-r2v-role2.mp4",
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/qpzxps/wan-r2v-object4.png",
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/wfjikw/wan-r2v-backgroud5.png"
    ]
  },
  "parameters": {
    "size": "1280*720",
    "duration": 10,
    "audio": true,
    "shot_type": "multi",
    "watermark": true
  }
}

Properties:

Name	Type	Required	Description
`model`	string	Yes	Model name. Available values: `wan2.6-r2v`, `wan2.6-r2v-flash`
`input`	object	Yes	Basic input such as the prompt and reference assets
`input.prompt`	string	Yes	Text prompt describing elements and visual traits of the generated video. See details below
`input.negative_prompt`	string	No	Negative prompt describing content to avoid. Supports Chinese and English. Maximum 500 characters; excess is truncated
`input.reference_urls`	array[string]	Yes	Array of reference file URLs. Supports videos and images. See details below
`parameters`	object	No	Video processing parameters such as resolution, duration, and shot type
`parameters.size`	string	No	Output video resolution in widthheight format (e.g. `1280720`); affects billing. See details below
`parameters.duration`	integer	No	Video duration in seconds (billed per second). Integer in `[2, 10]`. Default: `5`
`parameters.shot_type`	string	No	Shot type. `single` (default) for a single continuous shot; `multi` for multiple shots. Takes priority over the prompt
`parameters.audio`	boolean	No	Whether to generate audio. Supported only by `wan2.6-r2v-flash`. `true` (default) with audio; `false` silent. Pricing differs for audio vs silent
`parameters.watermark`	boolean	No	Whether to add a watermark (fixed text "AI生成" in the bottom-right corner). `false` (default) does not add it; `true` adds it
`parameters.seed`	integer	No	Random seed in `[0, 2147483647]`. Auto-generated if omitted. A fixed seed improves reproducibility but does not guarantee identical results

inputobject(Required)

Basic input such as the prompt and reference assets.

promptstring(Required)

Text prompt describing elements and visual traits of the generated video.

Supports Chinese and English. Each Chinese character, letter, or punctuation counts as one character; excess is truncated.

wan2.6-r2v, wan2.6-r2v-flash: maximum 1500 characters.

Character references: Reference roles with identifiers such as character1 and character2. Each reference (video or image) must contain only a single character. The model recognizes roles only through these identifiers. When multiple references are provided, role order follows the reference_urls array: the 1st URL is character1, the 2nd is character2, and so on.

Example: character1 happily watches a movie on the sofa.

reference_urlsarray[string](Required)

Array of uploaded reference file URLs. Supports videos and images. Used to extract character appearance and voice (if available) for consistent generation.

Each URL may point to one image or one video:

Images: 0–5
Videos: 0–3
Total limit: images + videos ≤ 5

When multiple references are provided, array order defines character order. Each reference file must contain only one subject character.

parametersobject(Optional)

Video processing parameters such as resolution, duration, and shot type.

sizestring(Optional)

Output video resolution in width*height format. Must be a concrete value (e.g. 1280*720), not 1:1 or 720P. Affects billing.

Default and available values depend on model:

wan2.6-r2v-flash, wan2.6-r2v: default 1920*1080 (1080P). Available resolutions: all 720P and 1080P options below.

720P tier:

Resolution	Aspect ratio
`1280*720`	16:9
`720*1280`	9:16
`960*960`	1:1
`1088*832`	4:3
`832*1088`	3:4

1080P tier:

Resolution	Aspect ratio
`1920*1080`	16:9
`1080*1920`	9:16
`1440*1440`	1:1
`1632*1248`	4:3
`1248*1632`	3:4

durationinteger(Optional)

Output video duration in seconds (billed per second).

wan2.6-r2v-flash, wan2.6-r2v: integer in [2, 10]. Default: 5

Example: 5

shot_typestring(Optional)

Specifies whether the video is a single continuous shot or multiple switching shots.

Priority: shot_type > prompt. For example, if shot_type is single, the model still outputs a single-shot video even if the prompt asks for multi-shot.

single: default, single continuous shot
multi: multiple shots

Use this when you need strict narrative control (e.g. single shot for product demos, multi-shot for short stories).

audioboolean(Optional)

Whether to generate audio. Supported only by wan2.6-r2v-flash. Pricing differs for audio vs silent.

true: default, output with audio
false: silent video. When generating silent video, you must explicitly set parameters.audio = false

Responses

200: Task created successfully

Content-Type: application/json

400: Invalid request parameters

Content-Type: application/json

429: Rate limit exceeded

Content-Type: application/json

Request Examples

curl -X POST "https://api.autorouter.top/api/v1/services/aigc/video-generation/video-synthesis" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-xxxxxx" \
  -H "X-DashScope-Async: enable" \
  -d '{
  "model": "wan2.6-r2v-flash",
  "input": {
    "prompt": "Character2 sits on a chair by the window, holding character3, and plays a gentle American country ballad next to character4. Character1 says to Character2: \"That sounds nice\"",
    "reference_urls": [
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/hfugmr/wan-r2v-role1.mp4",
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/qigswt/wan-r2v-role2.mp4",
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/qpzxps/wan-r2v-object4.png",
      "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20260129/wfjikw/wan-r2v-backgroud5.png"
    ]
  },
  "parameters": {
    "size": "1280*720",
    "duration": 10,
    "audio": true,
    "shot_type": "multi",
    "watermark": true
  }
}'

Response Examples

{
  "output": {
    "task_status": "PENDING",
    "task_id": "0385dc79-5ff8-4d82-bcb6-xxxxxx"
  },
  "request_id": "4909100c-7b5a-9f92-bfe5-xxxxxx"
}

Response fields:

Name	Type	Description
`output`	object	Task output
`output.task_id`	string	Task ID for polling status; valid for 24 hours
`output.task_status`	string	Task status. Enum: `PENDING`, `RUNNING`, `SUCCEEDED`, `FAILED`, `CANCELED`, `UNKNOWN`
`request_id`	string	Unique request ID for tracing and troubleshooting
`code`	string	Error code; returned only on failure
`message`	string	Error message; returned only on failure

GET `/api/v1/tasks/{task_id}`

Query Task Result by ID

Query task status and result using the task_id returned when creating the task. Query validity is 24 hours.

Headers

Name	Type	Required	Description
`Authorization`	string	Yes	Authentication. Format: `Bearer sk-xxxxxx`

Path Parameters

Name	Type	Required	Description
`task_id`	string	Yes	Task ID

Responses

200: Task queried successfully

Content-Type: application/json

400: Invalid request parameters

Content-Type: application/json

429: Rate limit exceeded

Content-Type: application/json

Request Examples

curl -X GET "https://api.autorouter.top/api/v1/tasks/{task_id}" \
  -H "Authorization: Bearer sk-xxxxxx"

Response Examples

{
  "request_id": "caa62a12-8841-41a6-8af2-xxxxxx",
  "output": {
    "task_id": "eff1443c-ccab-4676-aad3-xxxxxx",
    "task_status": "SUCCEEDED",
    "submit_time": "2025-12-16 00:25:59.869",
    "scheduled_time": "2025-12-16 00:25:59.900",
    "end_time": "2025-12-16 00:30:35.396",
    "orig_prompt": "character1 happily watches a movie on the sofa",
    "video_url": "https://dashscope-result-sh.oss-accelerate.aliyuncs.com/xxx.mp4?Expires=xxx"
  },
  "usage": {
    "duration": 10.0,
    "size": "1280*720",
    "input_video_duration": 5,
    "output_video_duration": 5,
    "video_count": 1,
    "SR": 720
  }
}

Response fields:

Name	Type	Description
`output`	object	Task output
`output.task_id`	string	Task ID; query validity 24 hours
`output.task_status`	string	Task status. Enum: `PENDING`, `RUNNING`, `SUCCEEDED`, `FAILED`, `CANCELED`, `UNKNOWN`. Typical flow while polling: `PENDING` → `RUNNING` → `SUCCEEDED` / `FAILED`
`output.submit_time`	string	Submit time, format `YYYY-MM-DD HH:mm:ss.SSS`
`output.scheduled_time`	string	Scheduled execution time, format `YYYY-MM-DD HH:mm:ss.SSS`
`output.end_time`	string	Completion time, format `YYYY-MM-DD HH:mm:ss.SSS`
`output.video_url`	string	Video download URL; returned only when `task_status` is `SUCCEEDED`. Link valid for 24 hours; video is MP4 (H.264)
`output.orig_prompt`	string	Original prompt corresponding to request parameter `prompt`
`output.code`	string	Error code; returned only on failure
`output.message`	string	Error message; returned only on failure
`usage`	object	Output statistics; returned only on success
`usage.input_video_duration`	integer	Input reference video duration in seconds
`usage.output_video_duration`	integer	Output video duration in seconds
`usage.duration`	float	Billable total duration in seconds: `input_video_duration + output_video_duration`
`usage.SR`	integer	Output resolution tier. Example: `720`
`usage.size`	string	Output resolution in widthheight format. Example: `1280720`
`usage.video_count`	integer	Number of output videos; always 1
`request_id`	string	Unique request ID for tracing and troubleshooting

Error Handling

HTTP 400 Parameter Errors (blocked before submit; no charge)

AutoRouter validates required fields before forwarding upstream:

Scenario	Response
Missing `input.prompt`	`{"code":"InvalidParameter","message":"...","request_id":"..."}`
Missing `input.reference_urls` or invalid asset count/combination	`{"code":"InvalidParameter","message":"...","request_id":"..."}`
Unknown model name	`{"code":"InvalidParameter","message":"unknown model: ...","request_id":"..."}`

HTTP 401 / 403 Auth Errors

401 Unauthorized: Invalid or expired API Key
403 Forbidden: API Key not allowed for this model (check token model whitelist)

HTTP 402 Insufficient Quota

Returns insufficient user quota. Please top up in the AutoRouter console.

Task `FAILED` Status

The task was submitted but upstream generation failed (output.task_status == "FAILED"). See output.code / output.message for details:

Cause	Suggestion
Content moderation failed	Adjust the prompt; avoid sensitive content
Media URL unreachable	Ensure image/video URLs are publicly accessible and not expired
Media file out of spec	Follow image/video limits in Request Body
Too many assets	Images + videos ≤ 5; max 3 videos, max 5 images
Invalid parameter combination (e.g. size not in list, duration out of range)	Follow Request Body rules
Silent video without disabling audio	For silent output with `wan2.6-r2v-flash`, set `parameters.audio = false`

When a task FAILED, AutoRouter automatically refunds to your account. Check refund records on the logs page.

Reference-to-Video

inputobject(Required)

Properties

Supported input formats

Reference video requirements

Reference image requirements

parametersobject(Optional)

Properties

On this page