| name | image-gen |
| description | Generate images using OpenAI's image generation API (gpt-image-1 or dall-e-3). Use whenever the user asks you to create, draw, design, or generate an image. After generating, send the image to the chat with the send_photo MCP tool. |
| allowed-tools | Bash(image-gen:*) |
Image Generator
Generate images via OpenAI's API and save them locally.
Quick start
image-gen "a cat sitting on a rainbow"
image-gen "a cat sitting on a rainbow" --model dall-e-3
image-gen "futuristic city skyline" --size 1536x1024
image-gen "app icon design" --size 1024x1024 --quality high
Usage
image-gen <prompt> [options]
Options
| Flag | Values | Default | Notes |
|---|
--model | gpt-image-1, dall-e-3 | gpt-image-1 | gpt-image-1 is newer and more capable |
--size | 1024x1024, 1536x1024, 1024x1536 | 1024x1024 | dall-e-3 also supports 1792x1024, 1024x1792 |
--quality | low, medium, high (gpt-image-1) or standard, hd (dall-e-3) | medium / standard | Higher quality = slower + more expensive |
--output | filename | auto-generated | Output path for the image |
Output
Prints the path to the saved image file. Use this path with the send_photo MCP tool to send it to the chat.
Workflow
- Generate the image:
image-gen "a watercolor painting of mountains at sunset"
- Send it to the user with the
send_photo tool using the output path.
Model comparison
- gpt-image-1: Latest model. Better prompt adherence, more photorealistic, supports text rendering in images. Slightly more expensive.
- dall-e-3: Previous generation. Still good for creative/artistic styles. Supports
1792x sizes.
Examples
image-gen "professional headshot photo, studio lighting, neutral background"
image-gen "impressionist painting of a rainy Paris street" --model dall-e-3
image-gen "minimal flat logo for a coffee shop called Bean There" --quality high
image-gen "infographic-style diagram of the solar system" --size 1536x1024
