Create Image Sets or Edit Reference Images · Alibaba

Describe a sequence of scenes to generate related images, or add up to nine references for editing. Choose 1K, 2K or supported custom dimensions; in set mode, the model chooses the actual image count within your cap.

Your AI

Use with your AI

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

CodexClaude CodeCursorWindsurfCline

View instructions

Use Toolmu generate-image-sets-fast-alibaba (standard Wan2.7 Image, ID 026d4eba-cde8-4fa4-960b-f700120fc93b) for a spring-windowsill image set. Reuse my existing connection; if setup is needed, follow https://toolmu.com/SKILL.md within my existing authorization and budget. Set body.prompt to: Create related scenes of a ceramic vase on a spring windowsill, aiming for a consistent vase and visual style. Show soft morning light, afternoon side light and warm evening light across the set, with no text. Set body.size=2K, body.enable_sequential=true and body.n=3. Omit body.images for this text-only task, and do not add a color palette or thinking mode to the set request. Three is an upper limit, not a guaranteed result count; do not select 4K. Complete this request only. On success, show the actual returned images and available download links, report the actual count and remind me to inspect subject details and save the downloads promptly. Further generation is a separate task within existing authorization and budget; do not add it automatically.

Task:
Create related pictures of [a ceramic vase on a spring windowsill]. Show [soft light in the morning], [sunlight from one side in the afternoon], and [warm light in the evening]. Try to keep [the vase and overall look] alike. I would like [up to three pictures], but fewer may be returned.

Paste into your connected AI's conversation.

Connection guide

Up to 0.432 credits / call

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

Alibaba

Create Image Sets or Edit Reference Images · Alibaba

You receive

Receive downloadable images after a successful request. For a scene series, ask for a ceramic vase on a spring windowsill in morning, afternoon and evening light. Set a cap of 1–12 images, then compare the actual returned scenes and save those you want to use.

Start with your task

Create related pictures of [a ceramic vase on a spring windowsill]. Show [soft light in the morning], [sunlight from one side in the afternoon], and [warm light in the evening]. Try to keep [the vase and overall look] alike. I would like [up to three pictures], but fewer may be returned.

What to know

This standard model does not support 4K. A set may contain fewer images than requested, and subject details, style continuity and lettering are not guaranteed. Processing time varies, and new model 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`
Requested count or set cap`body.n`	`integer`	Optional	Optional integer. Outside set mode, request 1–4 images, documented default 1. With enable_sequential=true, choose a cap of 1–12, documented default 12; the model chooses the actual count without exceeding the cap. Minimum `1` Maximum `12`
Random seed`body.seed`	`integer`	Optional	Optional integer from 0 to 2147483647. Reuses a random starting value; it does not ensure identical images or an unchanged subject. Minimum `0` Maximum `2147483647`
Image dimensions`body.size`	`string`	Optional	Optional 1K, 2K (default), or widthheight such as 15361024. All modes allow custom total pixel area from 768×768 to 2048×2048 and aspect ratio 1:8–8:1, without 4K. These are not independent side bounds; actual dimensions may vary slightly. Presets are square without references and use the last reference aspect ratio when images are supplied. Default `2K`
Ordered image references`body.images`	`array`	Optional	Optional array of up to nine public HTTPS image links for editing or multi-image reference. Keep the intended order. Omit the whole field for text-only generation. Maximum items `9`
Reference URL and file rules`body.images[]`	`string`	—	Nonempty public HTTPS URL. Each JPG/JPEG, PNG without alpha, BMP or WEBP image must be at most 20 MB, with both sides 240–8000 pixels and aspect ratio 1:8–8:1. Local files, base64 and HTTP links are not accepted reference inputs here. Minimum length `1`
Scene, edit or series description`body.prompt`	`string`	Required	Required nonempty text, at most 5000 characters. Describe a new image, an edit or the related scenes in a set. There is no separate negative_prompt input; include any exclusions in this description, without expecting guaranteed absence. Minimum length `1` Maximum length `5000`
Editing regions in source pixels`body.bbox_list`	`array`	Optional	Optional list aligned one-to-one with the input images. For each image, provide zero to two boxes [x1,y1,x2,y2] using four integer coordinates in original-image pixels, with top-left origin (0,0). Use [] for an image without boxes and choose coordinates within that image.
[]`body.bbox_list[]`	`array`	—
[]`body.bbox_list[][]`	`array`	—	Minimum items `4` Maximum items `4`
[]`body.bbox_list[][][]`	`integer`	—	Minimum `-9007199254740991` Maximum `9007199254740991`
AI watermark`body.watermark`	`boolean`	Optional	Optional boolean, default false. Enable it to request an AI Generated watermark. Default `false`
Custom color palette`body.color_palette`	`array`	Optional	Optional array of 3–10 colors; eight are recommended. Unavailable in set mode. Supply each color and its percentage share, with the shares totaling 100.00%. Minimum items `3` Maximum items `10`
[]`body.color_palette[]`	`object`	—	Required fields `hex, ratio`
Palette color`body.color_palette[].hex`	`string`	—	Required within each palette entry: a six-digit hexadecimal color preceded by #, such as #C2D1E6. Pattern `^#[0-9A-Fa-f]{6}$`
Color percentage`body.color_palette[].ratio`	`string`	—	Required percentage string within each palette entry. For provider compatibility, use two decimal places, such as 33.34%; all entries must add up to 100.00%. Pattern `^\d+(\.\d{1,2})?%$`
Thinking mode`body.thinking_mode`	`boolean`	Optional	Optional boolean, documented default true. Effective only for text-only generation outside set mode; it adds processing time without guaranteeing a better result.
Related image-set mode`body.enable_sequential`	`boolean`	Optional	Optional boolean, default false. Enable it to request related scenes in one set, with n acting as the count cap. Actual quantity and visual continuity are not guaranteed; thinking mode and the custom palette are unavailable. Default `false`

View raw JSON

{
  "type": "object",
  "required": [
    "body"
  ],
  "properties": {
    "body": {
      "type": "object",
      "$schema": "https://json-schema.org/draft/2020-12/schema",
      "required": [
        "prompt"
      ],
      "~standard": {
        "vendor": "zod",
        "version": 1,
        "jsonSchema": {}
      },
      "properties": {
        "n": {
          "type": "integer",
          "maximum": 12,
          "minimum": 1,
          "description": "Optional integer. Outside set mode, request 1–4 images, documented default 1. With enable_sequential=true, choose a cap of 1–12, documented default 12; the model chooses the actual count without exceeding the cap."
        },
        "seed": {
          "type": "integer",
          "maximum": 2147483647,
          "minimum": 0,
          "description": "Optional integer from 0 to 2147483647. Reuses a random starting value; it does not ensure identical images or an unchanged subject."
        },
        "size": {
          "anyOf": [
            {
              "enum": [
                "1K",
                "2K"
              ],
              "type": "string"
            },
            {
              "type": "string",
              "pattern": "^\\d+\\*\\d+$"
            }
          ],
          "default": "2K",
          "description": "Optional 1K, 2K (default), or width*height such as 1536*1024. All modes allow custom total pixel area from 768×768 to 2048×2048 and aspect ratio 1:8–8:1, without 4K. These are not independent side bounds; actual dimensions may vary slightly. Presets are square without references and use the last reference aspect ratio when images are supplied."
        },
        "images": {
          "type": "array",
          "items": {
            "type": "string",
            "minLength": 1,
            "description": "Nonempty public HTTPS URL. Each JPG/JPEG, PNG without alpha, BMP or WEBP image must be at most 20 MB, with both sides 240–8000 pixels and aspect ratio 1:8–8:1. Local files, base64 and HTTP links are not accepted reference inputs here."
          },
          "maxItems": 9,
          "description": "Optional array of up to nine public HTTPS image links for editing or multi-image reference. Keep the intended order. Omit the whole field for text-only generation."
        },
        "prompt": {
          "type": "string",
          "maxLength": 5000,
          "minLength": 1,
          "description": "Required nonempty text, at most 5000 characters. Describe a new image, an edit or the related scenes in a set. There is no separate negative_prompt input; include any exclusions in this description, without expecting guaranteed absence."
        },
        "bbox_list": {
          "type": "array",
          "items": {
            "type": "array",
            "items": {
              "type": "array",
              "items": {
                "type": "integer",
                "maximum": 9007199254740991,
                "minimum": -9007199254740991
              },
              "maxItems": 4,
              "minItems": 4
            }
          },
          "description": "Optional list aligned one-to-one with the input images. For each image, provide zero to two boxes [x1,y1,x2,y2] using four integer coordinates in original-image pixels, with top-left origin (0,0). Use [] for an image without boxes and choose coordinates within that image."
        },
        "watermark": {
          "type": "boolean",
          "default": false,
          "description": "Optional boolean, default false. Enable it to request an AI Generated watermark."
        },
        "color_palette": {
          "type": "array",
          "items": {
            "type": "object",
            "required": [
              "hex",
              "ratio"
            ],
            "properties": {
              "hex": {
                "type": "string",
                "pattern": "^#[0-9A-Fa-f]{6}$",
                "description": "Required within each palette entry: a six-digit hexadecimal color preceded by #, such as #C2D1E6."
              },
              "ratio": {
                "type": "string",
                "pattern": "^\\d+(\\.\\d{1,2})?%$",
                "description": "Required percentage string within each palette entry. For provider compatibility, use two decimal places, such as 33.34%; all entries must add up to 100.00%."
              }
            },
            "additionalProperties": false
          },
          "maxItems": 10,
          "minItems": 3,
          "description": "Optional array of 3–10 colors; eight are recommended. Unavailable in set mode. Supply each color and its percentage share, with the shares totaling 100.00%."
        },
        "thinking_mode": {
          "type": "boolean",
          "description": "Optional boolean, documented default true. Effective only for text-only generation outside set mode; it adds processing time without guaranteeing a better result."
        },
        "enable_sequential": {
          "type": "boolean",
          "default": false,
          "description": "Optional boolean, default false. Enable it to request related scenes in one set, with n acting as the count cap. Actual quantity and visual continuity are not guaranteed; thinking mode and the custom palette are unavailable."
        }
      },
      "additionalProperties": false
    }
  },
  "additionalProperties": false
}

Result schema

Output fields have not been documented yet.

Additional notes

To describe a scene series, start with the subject and visual style shared across the set, then explain the change in each scene. For the spring-windowsill example, morning softness, afternoon side light and evening warmth give the model a sequence to work with. A cap of three expresses the maximum number, so evaluate the scenes actually returned rather than assuming all three will appear.

Standard Wan2.7 Image also supports ordinary generation and reference-image editing. Use 1K, 2K or eligible custom dimensions; it does not offer 4K. For editing, provide eligible HTTPS references in order. Preset sizes without references are square, while references guide the preset aspect ratio through the last input. Custom dimensions remain subject to area and aspect rules. Download successful results promptly, and check details before using them.

Alibaba Wan image generation and editing documentation

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

Ordinary mode requests 1–4 images, with a documented default of one. Set mode requests related scenes with a cap of 1–12 images, default cap 12; the model chooses the actual count and may return fewer. Thinking mode and custom palettes are unavailable in set mode.

Create Image Sets or Edit Reference Images · Alibaba

Create Image Sets or Edit Reference Images · Alibaba

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Say what should stay the same

Say what should look different

Say how many pictures you want, then check them

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Say what should stay the same

Say what should look different

Say how many pictures you want, then check them

Questions about this tool.

How does set mode differ from ordinary multi-image generation?

Can a scene series be generated without reference images?

Can I specify exact width and height instead of a preset?