SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsimage-gen

image-gen

Design & creation

View repository

Updated 10-15 23:41

About this skill

GPT-image2

Call OpenAI's latest gpt-image-2 model to generate high-quality images. Just paste a prompt and it works.

Skill Features

  • Supports command-line batch generation, parameterized control of aspect ratio and resolution tier
  • Text-to-image + image-to-image dual mode,--image one parameter enables edit mode (up to 2 reference images)

Capabilities Overview

  • Text-to-image: Enter a prompt to generate a brand-new image
  • Image-to-image: Upload reference images (up to 2) + prompt, edit and generate based on the original image
  • Model: gpt-image-2 (OpenAI's latest image model)
  • API: RedFox new version gptImage2Submit / gptImage2Result
  • Output format: PNG (fixed output for the new API)
  • Aspect ratio：1:1 / 3:2 / 2:3 / 4:3 / 3:4 / 5:4 / 4:5 / 16:9(default) / 9:16 / 2:1 / 1:2 / 21:9 / 9:21
  • Resolution tier：1k / 2k(default) / 4k
  • Batch generation: Up to 4 images per request (new API limit)
  • Compatible with legacy pixel format: Still accepts 1792x1024 and other legacy formats; scripts automatically map them to aspect ratio + tier internally

Usage

Text-to-image — Enter text to generate an image

# 基本生成（默认 16:9 + 2k）
python3 "$SKILL_PATH/assets/imagegen.py" "一只橘色的猫咪坐在窗台上看着窗外的夕阳"

# 横版 4k 高清
python3 "$SKILL_PATH/assets/imagegen.py" "futuristic city skyline" --size 16:9 --resolution 4k

# 竖版小红书封面（3:4 + 2k）
python3 "$SKILL_PATH/assets/imagegen.py" "product cover, minimal style" --size 3:4 --resolution 2k

# 方形 1k 快速档
python3 "$SKILL_PATH/assets/imagegen.py" "minimalist cat logo, flat design" --size 1:1 --resolution 1k

# 批量生成 4 张（新接口上限）
python3 "$SKILL_PATH/assets/imagegen.py" "icon set, flat style" -n 4

# 兼容旧像素写法（自动映射为 16:9 + 1k）
python3 "$SKILL_PATH/assets/imagegen.py" "cyberpunk street" --size 1792x1024

Image-to-image — Upload reference images to edit and generate

# 单张参考图（自动上传 OSS → 提交任务）
python3 "$SKILL_PATH/assets/imagegen.py" "把猫咪改成白色，背景换成星空" --image ~/Pictures/cat.png

# 两张参考图（新接口最多支持 2 张）
python3 "$SKILL_PATH/assets/imagegen.py" "融合两张图的风格" --image ref1.png --image ref2.jpg

# 直接使用 URL 参考图（跳过上传步骤）
python3 "$SKILL_PATH/assets/imagegen.py" "把海报主体换成手表" --image "https://example.com/poster.jpg"

Other operations

# 仅提交任务（返回 taskId，不等待）
python3 "$SKILL_PATH/assets/imagegen.py" "complex scene" --no-download

# 查询已有任务结果
python3 "$SKILL_PATH/assets/imagegen.py" "" --task-id 5f100fcb8f3c4e3087c6aba93e121f7e

# 指定输出目录和文件名前缀
python3 "$SKILL_PATH/assets/imagegen.py" "illustration" -o ~/Pictures/AI --prefix artwork

Parameter Description

Parameter Description Default value
prompt Generation/edit prompt (required, up to 500 characters) -
--size Aspect ratio (e.g. 16:9); also compatible with legacy pixel formats (e.g. 1792x1024） 16:9
--resolution Resolution tier:1k / 2k / 4k Pixel format is automatically matched; aspect ratio defaults to 2k
-n, --count Number of images to generate (1-4, new API limit is 4) 1
--image Reference image path or URL (can be passed multiple times, up to 2) -
-o, --output-dir Output directory ~/Downloads/QoderImages
--prefix Filename prefix image
--no-download Submit only without waiting -
--task-id Query an existing task -
--api-key Specify API Key -

Deprecated parameters (no longer supported by the new API; if passed, they will be ignored with a notice)

Parameter Description
--quality Use --resolution instead
--format The new API has fixed PNG output
--bg / --background The new API no longer supports background control
--compression The new API no longer supports compression ratio
--fidelity The new API no longer supports fidelity control

Dependency Installation

Dependency Installation command
requests pip3 install requests

First Use

Configure the API Key first, then run:

# 设置环境变量
export REDFOX_API_KEY=ak_你的密钥

# 运行
python3 "$SKILL_PATH/assets/imagegen.py" "一只橘色的猫咪"

Go to redfox.hk to register and get an API Key.

⚠️ The new API only supportspaid calls, and free account credits cannot be used for this API. If error code 3203is returned, please go to the top-up page to add paid credits.

Subsequent Use

Go to redfox.hk to register an account and get your own API Token. Choose any one of the three configuration methods:

Configuration method Description Command
Environment variables(Recommended) Set once, takes effect globally export REDFOX_API_KEY=ak_你的密钥
Command-line arguments Temporary use, takes effect for a single run python3 "$SKILL_PATH/assets/imagegen.py" "prompt" --api-key ak_你的密钥
Configuration file Persistent storage, retained across sessions mkdir -p ~/.qoder/apis && echo '{"api_key":"ak_你的密钥"}' > ~/.qoder/apis/redfox.json

API specification (new)

Submit task

POST https://redfox.hk/story/api/parseWork/imageGen/gptImage2Submit

Request headers:REDFOX_API_KEY: ak_xxx + Content-Type: application/json

Request body:

{
  "prompt": "把这个海报的主体变为手表 并把文字都用中文",
  "resolution": "2k",
  "size": "16:9",
  "n": 2,
  "referenceImages": ["https://example.com/poster.jpg"]
}

Response:data.taskId used for subsequent polling.

Query result

POST https://redfox.hk/story/api/parseWork/imageGen/gptImage2Result

Request body:{"taskId": "..."}

Key response fields:

Field Description
data.status completed / processing / queued / failed
data.progress Generation progress 0-100
data.imageUrls Array of generated result URLs (count matches n consistent)
data.failReason Failure reason (null on success)
data.model Model used (gpt-image-2）
data.resolution / data.size Actual tier and aspect ratio used

FAQ

Q: What are the features of this Skill? A: Directly call the gpt-image-2 model from the command line, supporting batch generation, aspect ratio and resolution tier control, and image-to-image (up to 2 reference images).

Q: How long does it take to generate an image? A: Usually 10-60 seconds,4k tier or complex scenes may take longer. The script automatically polls and waits and displays progress.

Q: What is the difference between the new resolution and the old quality ? A：resolution is a resolution tier (1k/2k/4k), which directly determines the clarity of the output image and the generation time; the old quality parameter is deprecated and will be ignored if passed in.

Q: Why was --size changed from pixels to aspect ratio? A: The new API gptImage2Submit 's size field is the aspect ratio (e.g. 16:9). For compatibility with old calls, the script still accepts 1792x1024 and other pixel formats, and internally maps them automatically to aspect ratio + recommended tier.

Q: How many reference images can be passed for image-to-image? A: The new API supports up to 2. Just pass --image multiple times; any excess will be truncated with a prompt.

Q: Why does the call return error code 3203? A: The new API only supports paid calls; free account credits cannot be used. Please go to redfox.hk/dashboard/recharge to top up paid credits.

Q: How do I get an API Key? A: Go to redfox.hk to register and get your own API Token.

Q: Which image formats are supported as reference images? A: Local files or HTTP(S) URLs in PNG, JPEG, and WebP formats are supported.

Q: Is there a length limit for prompts? A: Prompts support up to 500 characters; anything over that will be blocked with a prompt to shorten it.