SuperAPI AI Directory

jimeng-ai

General & other

View repository

Updated 10-16 02:35

About this skill

Jimeng AI Text-to-Image/Text-to-Video Skill

Based on the text-to-image and text-to-video capabilities of Volcano Engine Jimeng AI, it supports generating images and videos from text descriptions.

Features

  • Execution process storage: Uses MD5(prompt) as the folder name to save task status
  • Asynchronous query: Supports resumable execution to avoid resubmitting the same task
  • Base64 image processing: Decodes and saves images directly from the API response
  • Supports Jimeng AI text-to-image (v3.0 / v3.1 / v4.0)
  • Supports Jimeng AI text-to-video (v3.0 1080P)
  • Configurable aspect ratio, generation count, and custom dimensions

Required Environment Variables

Variable Name Required Description
VOLCENGINE_AK Required Volcano Engine Access Key
VOLCENGINE_SK Conditionally required Volcano Engine Secret Key (required for permanent credentials)
VOLCENGINE_TOKEN Optional Security token (required for temporary credentials STS)

Note: When using temporary credentials (starting with AKTP), you can use only AK + Token, and SK is not required.

Configuration Example

export VOLCENGINE_AK="your-access-key"
export VOLCENGINE_SK="your-secret-key"

# 如果使用临时凭证(STS),还需要设置 Token
export VOLCENGINE_TOKEN="your-security-token"

How to obtain:

  1. Log in to Volcano Engine Console
  2. Go to "Access Control" -> "Key Management"
  3. Create or view an existing access key

Install Dependencies

cd ~/.openclaw/workspace/skills/jimeng-ai
npm install

Workflow

First Execution (New Task)

When running with a new prompt, the script will:

  1. Submit the task to the API
  2. Use md5(提示词) as the folder name to create a directory
  3. Save param.json、response.json and taskId.txt
  4. Output:"任务已提交,TaskId: xxx"
$ npx ts-node scripts/text2image.ts "一只可爱的猫咪"
任务已提交,TaskId: 1234567890

Subsequent Execution (Asynchronous Query)

Running with the same prompt will query the existing task:

  1. If the image already exists -> return the image path immediately
  2. If the task is not completed -> output:"任务未完成,TaskId: xxx"
  3. If the task is completed -> from binary_data_base64 decode and save the image
$ npx ts-node scripts/text2image.ts "一只可爱的猫咪"
任务未完成,TaskId: 1234567890

# 或者任务完成时:
$ npx ts-node scripts/text2image.ts "一只可爱的猫咪"
任务已完成,图片保存路径:
  - ./output/<md5_hash>/1.jpg
  - ./output/<md5_hash>/2.jpg

Usage

Basic Usage

npx ts-node scripts/text2image.ts "一只可爱的猫咪"

Full Parameters

npx ts-node scripts/text2image.ts "提示词" \
  --version v40 \
  --ratio 16:9 \
  --count 2

Parameter Description

Parameter Description Default Value
prompt Image generation prompt (required) -
--version API version: v30, v31, v40 v31
--ratio Aspect ratio: 1:1, 9:16, 16:9, 3:4, 4:3, 2:3, 3:2, 1:2, 2:1 16:9
--count Generation count 1-4 1
--width Specify width (optional) -
--height Specify height (optional) -
--size Specify area (optional, e.g. 4194304 means 2048x2048) -
--seed Random seed (optional) -
--output Image output directory ./output
--debug Debug mode false
--no-download Do not download the image, only return the URL false

Output Format

Task Submitted (First Run)

{
  "success": true,
  "submitted": true,
  "prompt": "一只可爱的猫咪",
  "version": "v40",
  "ratio": "1:1",
  "count": 1,
  "taskId": "1234567890",
  "folder": "./output/<md5_hash>",
  "message": "任务已提交,请稍后使用相同提示词查询结果"
}

Task Completed

{
  "success": true,
  "prompt": "一只可爱的猫咪",
  "version": "v40",
  "ratio": "1:1",
  "count": 1,
  "taskId": "1234567890",
  "images": [
    "./output/<md5_hash>/1.jpg",
    "./output/<md5_hash>/2.jpg"
  ],
  "outputDir": "./output/<md5_hash>"
}

Task Not Completed

{
  "success": true,
  "prompt": "一只可爱的猫咪",
  "version": "v40",
  "ratio": "1:1",
  "count": 1,
  "taskId": "1234567890",
  "folder": "./output/<md5_hash>",
  "message": "任务未完成,请稍后使用相同提示词查询结果"
}

Error Response

{
  "success": false,
  "error": {
    "code": "MISSING_CREDENTIALS",
    "message": "请设置环境变量 VOLCENGINE_AK 和 VOLCENGINE_SK"
  }
}

Folder Structure

output/
└── <md5(prompt)>/           # md5哈希作为文件夹名
    ├── param.json           # 请求参数
    ├── response.json        # API提交响应
    ├── taskId.txt           # 任务ID
    └── 1.jpg, 2.jpg, ...    # 生成的图片

Examples

Generate a landscape painting

npx ts-node scripts/text2image.ts "山水风景画,水墨风格" --version v40 --ratio 16:9

Generate a sci-fi city

npx ts-node scripts/text2image.ts "未来科幻城市,霓虹灯光,赛博朋克风格" --version v40 --ratio 16:9 --count 2

Generate with specified dimensions

npx ts-node scripts/text2image.ts "抽象艺术" --width 2048 --height 1152

Custom output directory

npx ts-node scripts/text2image.ts "一只可爱的猫咪" --output ~/Pictures/jimeng

Version Notes

  • v30: Jimeng 3.0 Basic Version
  • v31: Jimeng 3.1 Improved Version
  • v40: Jimeng 4.0 Latest Version (Recommended)

Text-to-Video Usage

Basic Usage

npx ts-node scripts/text2video.ts "一只可爱的猫咪在草地上奔跑"

Full Parameters

npx ts-node scripts/text2video.ts "提示词" \
  --ratio 9:16 \
  --duration 5 \
  --fps 24

Parameter Description

Parameter Description Default Value
prompt Video generation prompt (required) -
--ratio Aspect ratio: 16:9, 4:3, 1:1, 3:4, 9:16, 21:9 9:16
--duration Video duration: 5 or 10 seconds 5
--fps Frame rate: 24 or 30 24
--output Video output directory ./output
--wait Wait for task completion false
--debug Debug mode false
--no-download Do not download the video, only return the URL false

Video output format

Task submitted

{
  "success": true,
  "submitted": true,
  "prompt": "元宵节灯笼",
  "ratio": "9:16",
  "duration": 5,
  "fps": 24,
  "taskId": "1234567890",
  "folder": "./output/video/<md5_hash>",
  "message": "任务已提交,请稍后使用相同提示词查询结果"
}

Task completed

{
  "success": true,
  "prompt": "元宵节灯笼",
  "ratio": "9:16",
  "duration": 5,
  "fps": 24,
  "taskId": "1234567890",
  "videoUrl": "https://...",
  "data": {}
}

Task not completed

{
  "success": true,
  "pending": true,
  "prompt": "元宵节灯笼",
  "ratio": "9:16",
  "duration": 5,
  "fps": 24,
  "taskId": "1234567890",
  "status": "in_queue",
  "message": "任务处理中,请稍后使用相同提示词查询结果"
}

Video folder structure

output/video/
└── <md5(prompt)>/           # md5哈希作为文件夹名
    ├── param.json           # 请求参数
    ├── response.json        # API提交响应
    ├── taskId.txt           # 任务ID
    └── video.mp4            # 生成的视频

Reference Documentation