Create a New Scene from Character and Setting References · Alibaba

Describe a scene and provide reference images or character clips. Wan2.7 uses them to create a new short video; optional voice references guide how a character sounds.

Your AI

Use with your AI

For tasks with several steps, continue in the AI you already use.

CodexClaude CodeCursorWindsurfCline

View instructions

Use Toolmu reference-to-video-alibaba (Alibaba Wan2.7 reference-to-video; ID 0ff49d37-382a-4896-81da-3d7b60ed1e19). Reuse my connection; if setup is needed, follow https://toolmu.com/SKILL.md within my existing authorization and budget. First ask for two real eligible public HTTPS image URLs: one illustration containing a single character, then one room-background image. Do not invent or substitute assets. Put both in body.media as reference_image, preserving that order as Image 1 and Image 2. Set body.prompt to The character from Image 1 walks across the room from Image 2. Set body.duration=5, body.resolution=720P, body.ratio=16:9 and body.prompt_extend=true. Supply no reference_voice or first_frame. Generate only this requested scene. After actual successful processing, show the returned MP4 and available download link. Remind me to inspect the character, room, movement and sound rather than assume exact fidelity, and save the temporary video promptly. Submission alone is not a completed scene.

Task:
Create [a five-second video] of [my character] walking through [my room]. I will share [a link anyone can open to a picture of my character without signing in] and [a link anyone can open to the room picture without signing in]. Show me the finished video.

Paste into your connected AI's conversation.

Connection guide

Up to 2.7 credits / call

The final charge is based on confirmed usage. Unused reserved credits return to your available balance.

Alibaba

Create a New Scene from Character and Setting References · Alibaba

Describe a scene and provide reference images or character clips. Wan2.7 uses them to create a new short video; optional voice references guide how a character sounds.

You receive

Get one H.264 MP4 at 720P or 1080P. For this example, first supply your real public HTTPS single-character illustration and room-background image, then request that character walking through the room for five seconds. Review the actual appearance, setting, movement and sound before saving.

Start with your task

Create [a five-second video] of [my character] walking through [my room]. I will share [a link anyone can open to a picture of my character without signing in] and [a link anyone can open to the room picture without signing in]. Show me the finished video.

What to know

Use at least one reference image or video, with at most five combined. Character references must each show one character. A reference video limits output to 2–10 seconds; image-only requests allow 2–15. Appearance, voice and scene fidelity are not guaranteed. Submission is not a finished clip, processing time varies and new download links typically last about 24 hours.

Inputs & outputs

Inspect the current inputs, availability, and maximum charge before starting.

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
body`body`	`object`	Required	Required fields `prompt, media`
Random seed`body.seed`	`integer`	Optional	Optional integer from 0 to 2147483647. A fixed seed reuses a random starting point, but does not ensure the same scene, character appearance or voice. Minimum `0` Maximum `2147483647`
Scene references`body.media`	`array`	Required	Required nonempty array containing at least one reference_image or reference_video. Those two types may total at most five entries. You may add one first_frame separately, but a first frame alone does not satisfy the reference requirement. Keep the order: images and videos are numbered separately as Image 1 and Video 1 in the prompt. Minimum items `1`
[]`body.media[]`	`object`	—	Required fields `type, url`
Reference asset URL`body.media[].url`	`string`	—	Required public HTTPS URL for the image or video. HTTP, Base64 and local-file inputs are not accepted. Images must be JPG/JPEG, PNG without alpha, BMP or WEBP, at most 20 MB, each side 240–8000 pixels and aspect ratio 1:8–8:1. Videos must be MP4 or MOV, 1–30 seconds, at most 100 MB, each side 240–4096 pixels and aspect ratio 1:8–8:1. Minimum length `1`
How this material is used`body.media[].type`	`string`	—	Required value: reference_image for a character or setting, reference_video for a character, or first_frame for the opening composition. Each character reference must contain one character, which may be a person, animal or object; a background image can provide the scene. Avoid empty-scene videos as character references. References guide generation without guaranteeing exact likeness. Options `reference_image, reference_video, first_frame`
Character voice reference`body.media[].reference_voice`	`string`	—	Optional nonempty public HTTPS audio URL, only on reference_image or reference_video, not first_frame. Use WAV or MP3, 1–10 seconds and at most 15 MB. It guides voice characteristics, not spoken content. A supplied voice takes priority over reference-video audio; otherwise available video audio serves as its voice reference. Voice likeness, exact dialogue and lip-sync are not guaranteed. Minimum length `1`
Scene aspect ratio`body.ratio`	`string`	Optional	Optional 16:9, 9:16, 1:1, 4:3 or 3:4, default 16:9. When a first_frame is supplied, this setting is ignored and its image framing is used instead. It does not guarantee precise composition or character fidelity. Default `16:9` Options `16:9, 9:16, 1:1, 4:3, 3:4`
Scene and reference instructions`body.prompt`	`string`	Required	Required nonempty text, at most 5000 characters. Describe the new action or dialogue and identify references with capitalized, spaced names such as Image 1 and Video 1. Number images and videos separately in their array order. A voice sample does not supply the words to be spoken; describe desired dialogue here. Minimum length `1` Maximum length `5000`
New video duration`body.duration`	`integer`	Optional	Optional whole seconds, default 5. Image-only requests allow 2–15; if any reference_video is included, use 2–10. This is the newly generated scene length, not the original clip plus a continuation. Default `5` Minimum `2` Maximum `15`
AI-generated watermark`body.watermark`	`boolean`	Optional	Optional boolean, default false. Set true to request the AI Generated watermark on the video. Default `false`
Output resolution`body.resolution`	`string`	Optional	Optional 720P or 1080P, default 720P for this Tool. Select a tier for the generated video; resolution does not ensure sharper details or closer adherence to references. Default `720P` Options `720P, 1080P`
Rewrite the scene description`body.prompt_extend`	`boolean`	Optional	Optional boolean, default true. It allows the model to rewrite the prompt before generation and can increase processing time. Rewriting does not guarantee higher quality or faithful interpretation. Default `true`
Unwanted visual elements`body.negative_prompt`	`string`	Optional	Optional text, at most 500 characters, to guide exclusions. It does not guarantee that specified elements will be absent from the video. Maximum length `500`

View raw JSON

{
  "type": "object",
  "required": [
    "body"
  ],
  "properties": {
    "body": {
      "type": "object",
      "$schema": "https://json-schema.org/draft/2020-12/schema",
      "required": [
        "prompt",
        "media"
      ],
      "~standard": {
        "vendor": "zod",
        "version": 1,
        "jsonSchema": {}
      },
      "properties": {
        "seed": {
          "type": "integer",
          "maximum": 2147483647,
          "minimum": 0,
          "description": "Optional integer from 0 to 2147483647. A fixed seed reuses a random starting point, but does not ensure the same scene, character appearance or voice."
        },
        "media": {
          "type": "array",
          "items": {
            "type": "object",
            "required": [
              "type",
              "url"
            ],
            "properties": {
              "url": {
                "type": "string",
                "minLength": 1,
                "description": "Required public HTTPS URL for the image or video. HTTP, Base64 and local-file inputs are not accepted. Images must be JPG/JPEG, PNG without alpha, BMP or WEBP, at most 20 MB, each side 240–8000 pixels and aspect ratio 1:8–8:1. Videos must be MP4 or MOV, 1–30 seconds, at most 100 MB, each side 240–4096 pixels and aspect ratio 1:8–8:1."
              },
              "type": {
                "enum": [
                  "reference_image",
                  "reference_video",
                  "first_frame"
                ],
                "type": "string",
                "description": "Required value: reference_image for a character or setting, reference_video for a character, or first_frame for the opening composition. Each character reference must contain one character, which may be a person, animal or object; a background image can provide the scene. Avoid empty-scene videos as character references. References guide generation without guaranteeing exact likeness."
              },
              "reference_voice": {
                "type": "string",
                "minLength": 1,
                "description": "Optional nonempty public HTTPS audio URL, only on reference_image or reference_video, not first_frame. Use WAV or MP3, 1–10 seconds and at most 15 MB. It guides voice characteristics, not spoken content. A supplied voice takes priority over reference-video audio; otherwise available video audio serves as its voice reference. Voice likeness, exact dialogue and lip-sync are not guaranteed."
              }
            },
            "additionalProperties": false
          },
          "minItems": 1,
          "description": "Required nonempty array containing at least one reference_image or reference_video. Those two types may total at most five entries. You may add one first_frame separately, but a first frame alone does not satisfy the reference requirement. Keep the order: images and videos are numbered separately as Image 1 and Video 1 in the prompt."
        },
        "ratio": {
          "enum": [
            "16:9",
            "9:16",
            "1:1",
            "4:3",
            "3:4"
          ],
          "type": "string",
          "default": "16:9",
          "description": "Optional 16:9, 9:16, 1:1, 4:3 or 3:4, default 16:9. When a first_frame is supplied, this setting is ignored and its image framing is used instead. It does not guarantee precise composition or character fidelity."
        },
        "prompt": {
          "type": "string",
          "maxLength": 5000,
          "minLength": 1,
          "description": "Required nonempty text, at most 5000 characters. Describe the new action or dialogue and identify references with capitalized, spaced names such as Image 1 and Video 1. Number images and videos separately in their array order. A voice sample does not supply the words to be spoken; describe desired dialogue here."
        },
        "duration": {
          "type": "integer",
          "default": 5,
          "maximum": 15,
          "minimum": 2,
          "description": "Optional whole seconds, default 5. Image-only requests allow 2–15; if any reference_video is included, use 2–10. This is the newly generated scene length, not the original clip plus a continuation."
        },
        "watermark": {
          "type": "boolean",
          "default": false,
          "description": "Optional boolean, default false. Set true to request the AI Generated watermark on the video."
        },
        "resolution": {
          "enum": [
            "720P",
            "1080P"
          ],
          "type": "string",
          "default": "720P",
          "description": "Optional 720P or 1080P, default 720P for this Tool. Select a tier for the generated video; resolution does not ensure sharper details or closer adherence to references."
        },
        "prompt_extend": {
          "type": "boolean",
          "default": true,
          "description": "Optional boolean, default true. It allows the model to rewrite the prompt before generation and can increase processing time. Rewriting does not guarantee higher quality or faithful interpretation."
        },
        "negative_prompt": {
          "type": "string",
          "maxLength": 500,
          "description": "Optional text, at most 500 characters, to guide exclusions. It does not guarantee that specified elements will be absent from the video."
        }
      },
      "additionalProperties": false
    }
  },
  "additionalProperties": false
}

Result schema

Output fields have not been documented yet.

Additional notes

Give each reference a clear purpose. In this example, one image defines the character and another supplies the room; naming both in the action tells the model which material should inform each part of the scene. Check that the character reference contains only one subject.

For mixed image and video references, count each type separately in the order supplied so the written instructions identify the intended asset. A simple action makes it easier to compare the finished scene with your references. See Alibaba reference-to-video documentation.

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

It guides a character’s voice characteristics, not the words to be spoken. Describe desired dialogue in the prompt. A separate voice reference overrides the reference video’s audio as the voice source; without one, available video audio can be used as the reference. Neither exact dialogue nor original-track preservation is guaranteed.

Create a New Scene from Character and Setting References · Alibaba

Create a New Scene from Character and Setting References · Alibaba

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Share a picture with one character

Share a picture of the room

Check how the scene turned out

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Share a picture with one character

Share a picture of the room

Check how the scene turned out

Questions about this tool.

Does a voice reference provide the dialogue or preserve the original soundtrack?

Can a 30-second reference video produce a 30-second result?

What changes when I add a first frame?