Skip to content

Image Generation

Generate stunning images using OpenAI's gpt-image-1 through the MCP interface.

Overview

The generate-image tool provides access to gpt-image-1's advanced image generation capabilities featuring:

  • Advanced Text Rendering: Crisp, legible typography and logos
  • Native Transparency: Built-in transparent background support
  • Superior Instruction Following: Inherits GPT-4o's language understanding
  • Multiple Output Formats: PNG, JPEG, WebP with optimized compression
  • Flexible Dimensions: Optimized aspect ratios for different use cases

Aspect Ratios

Choose from three aspect ratios optimized for different use cases:

Square (1:1)

  • Resolution: 1024x1024
  • Perfect for: Icons, avatars, social media posts
  • Use case: Profile pictures, app icons, thumbnails

Landscape (16:9)

  • Resolution: 1536x1024 (optimized for gpt-image-1)
  • Perfect for: Headers, banners, presentations
  • Use case: Website heroes, blog headers, desktop wallpapers

Portrait (9:16)

  • Resolution: 1024x1536 (optimized for gpt-image-1)
  • Perfect for: Mobile screens, stories, posters
  • Use case: Phone wallpapers, Instagram stories, book covers

Quality Options

Standard Quality

  • Faster generation time
  • Lower token cost
  • Suitable for drafts and iterations

HD Quality

  • Maximum detail and clarity
  • Longer generation time
  • Best for final production images

Style Variations

Vivid Style

  • Enhanced colors and contrast
  • More artistic interpretation
  • Dramatic and eye-catching results

Natural Style

  • Realistic rendering
  • Subtle colors and lighting
  • Photographic quality

Prompt Engineering

Best Practices

  1. Be Specific: Include details about composition, lighting, and mood
  2. Use Art Terms: Reference artistic styles, techniques, or movements
  3. Specify Medium: Mention if you want photography, illustration, 3D render, etc.
  4. Include Context: Describe the environment and atmosphere

Example Prompts

Product Photography

A minimalist product shot of a sleek smartphone on a white marble surface,
soft studio lighting, shallow depth of field, professional photography

Illustration

A whimsical children's book illustration of a friendly dragon reading
stories to woodland creatures, watercolor style, warm autumn colors

Concept Art

Futuristic cityscape at sunset, cyberpunk aesthetic, neon lights
reflecting on wet streets, flying vehicles, digital art style

File Management

Naming Strategies

  • timestamp: image_20250114_123456.png
  • prompt: serene_japanese_garden.png
  • custom: Your specified filename
  • hash: img_a1b2c3d4.png

Organization Options

  • none: All images in one directory
  • date: Organized by date (2025-01-14/)
  • aspect_ratio: Grouped by ratio (landscape/)
  • quality: Separated by quality (hd/)

Advanced Features

Image Editing

Edit and enhance generated images:

typescript
// Generate base image
const baseImage = await client.callTool("generate-image", {
  prompt: "A mountain landscape",
  aspect_ratio: "landscape",
});

// Edit the image
const editedImage = await client.callTool("edit-image", {
  source_image: baseImage.data.file_path,
  edit_prompt: "Add dramatic storm clouds",
  edit_type: "background_change",
});

Batch Editing

Apply the same edit to multiple images:

typescript
const result = await client.callTool("batch-edit", {
  image_urls: ["image1.png", "image2.png", "image3.png"],
  edit_prompt: "Convert to black and white",
  edit_type: "style_transfer",
});

Tips for Success

  1. Iterate Quickly: Start with standard quality for experimentation
  2. Save Versions: Keep track of successful prompts
  3. Combine Styles: Mix artistic references for unique results
  4. Use Negatives: Specify what you don't want in the image
  5. Test Variations: Try different aspect ratios for the same concept

Limitations

  • Maximum prompt length: 4000 characters
  • No explicit content generation
  • Some artistic styles may be interpreted differently
  • Generation time varies based on server load

Next Steps

  • Learn about LLM translation patterns for multilingual support
  • Explore Image Editing capabilities for post-processing
  • Try Batch Editing for processing multiple images
  • Check out Examples for more use cases

Released under the MIT License.