Wan 2.5 Text-To-Image Preview Generation

Generate images from text prompts using the wan2.5-t2i-preview model.

POST /vendors/alibaba/v1/wan2.5-t2i-preview/generation

Create Generation Task

cURL

curl --request POST \
  --url https://api.mulerouter.ai/vendors/alibaba/v1/wan2.5-t2i-preview/generation \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "prompt": "<string>",
  "negative_prompt": "<string>",
  "size": "1280*1280",
  "n": 4,
  "prompt_extend": false,
  "seed": 1073741823,
  "safety_filter": true
}
'

202

{
  "task_info": {
    "id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
    "created_at": "2023-11-07T05:31:56Z",
    "updated_at": "2023-11-07T05:31:56Z"
  }
}

This API is compatible with Alibaba’s Tongyi Wanxiang (Wan2) models. Please refer to Alibaba Cloud’s official documentation for more details.

Overview

Generate high-quality images from text descriptions using the wan2.5-t2i-preview model. This is the latest Wan2.5 text-to-image model with enhanced capabilities.

Key Features

• Support for longer prompts (up to 2000 characters)
• Flexible resolution and aspect ratio support
• Higher quality output

Resolution Guidelines

• Default: 1280*1280
• Total pixels: Between 768*768 and 1440*1440
• Aspect ratio: Between 1:4 and 4:1
• Example: 768*2700 is valid (meets pixel and ratio requirements)

Common Aspect Ratios

Aspect Ratio	Recommended Resolution
1:1	1280*1280 or 1024*1024
2:3	800*1200
3:2	1200*800
3:4	960*1280
4:3	1280*960
9:16	720*1280
16:9	1280*720
21:9	1344*576

Prompt Tips

For best results:

• Be specific and descriptive about what you want to see
• Include details about style, composition, lighting, and mood
• Use the negative prompt to exclude unwanted elements
• For short prompts, consider enabling prompt_extend for enhanced results

Example Requests

Basic Text-to-Image

{
  "prompt": "A sitting orange cat with a joyful expression, lively and cute, realistic and accurate",
  "n": 1
}

With Negative Prompt

{
  "prompt": "A serene mountain landscape at sunset, dramatic clouds, vibrant colors",
  "negative_prompt": "low resolution, error, worst quality, low quality, incomplete, extra fingers, poor proportions",
  "size": "1280*720",
  "n": 4
}

High-Resolution Portrait

{
  "prompt": "Professional portrait of a woman in business attire, studio lighting, shallow depth of field, professional photography",
  "size": "960*1280",
  "n": 1,
  "prompt_extend": false
}

With Custom Seed for Reproducibility

{
  "prompt": "Futuristic cityscape at night, neon lights, cyberpunk style",
  "size": "1344*576",
  "n": 1,
  "seed": 12345
}

Parameters

n (Number of Images)

• Range: 1-4
• Default: 4
• Note: Directly affects cost (cost = unit price × number of images)
• Recommendation: Set to 1 during testing to minimize costs

prompt_extend (Prompt Rewriting)

• Default: false
• Effect: Uses AI to enhance and expand short prompts
• Trade-off: Improves results for short prompts but increases processing time

seed (Random Seed)

• Range: 0 to 2,147,483,647
• Purpose: Improves reproducibility of results
• Note: Due to probabilistic nature, same seed doesn’t guarantee identical results
• Behavior: When n > 1, generates seeds as: seed, seed+1, seed+2, etc.

Authorizations

Authorization · string · header · required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

prompt · string · required

Text description for the image (max 2000 characters).

Maximum string length: 2000

negative_prompt · string

Negative prompt describing unwanted content (max 500 characters)

Maximum string length: 500

size · string · default: 1280*1280

Image resolution in format "width*height" (e.g., "1280*1280")

Default: 1280*1280 Total pixels: [768*768, 1440*1440], aspect ratio: [1:4, 4:1]

n · integer · default: 4

Number of images to generate (1-4)

Required range: 1 <= x <= 4

prompt_extend · boolean · default: false

Enable intelligent prompt rewriting (improves short prompts but increases processing time)

seed · integer

Random seed for reproducibility [0, 2147483647]

Required range: 0 <= x <= 2147483647

safety_filter · boolean | null · default: true

Enable content safety filter. Defaults to true. Set to false to disable content safety inspection.

Response

202 - application/json

Accepted - Task created successfully

task_info · object

Show child attributes