A Python library for programmatic poster and image design with optional MCP (Model Context Protocol) integration for AI-driven design workflows.

• Simple Python API: Create posters with a fluent, chainable API
• Multiple Element Types: Text, images, and shapes with full customization
• Preset Templates: Ready-to-use canvas sizes for social media (Instagram, Facebook, WeChat)
• Layout System: Automatic alignment, distribution, and grid layouts
• Style Management: Themes and color schemes for consistent designs
• Image Filters: Blur, brightness, contrast, and more
• MCP Integration: Optional AI-driven design through natural language
• Session Management: Multi-user support for production deployments

pip install poster-design

pip install poster-design[mcp]

git clone https://github.com/yourusername/poster-design.git
cd poster-design
pip install -e ".[dev]"

from poster_design import Canvas, TextElement, ImageElement

# Create canvas with preset
canvas = Canvas(preset="instagram_post")

# Set background using different options
# Option 1: Preset background
canvas.set_background(preset="paper_texture")

# Option 2: Gradient background
canvas.set_background(gradient=["#FFB6C1", "#FFA07A"])

# Option 3: Solid color
canvas.set_background(color="#FF6B6B")

# Add title
title = TextElement(
    text="Summer Sale",
    font_size=72,
    color="#FFFFFF",
    bold=True
)
canvas.add(title, position="center", offset_y=-50)

# Add subtitle
subtitle = TextElement(
    text="50% OFF",
    font_size=96,
    color="#FF6347",
    bold=True
)
canvas.add(subtitle, position="center", offset_y=50)

# Export
canvas.export("summer_sale.png", dpi=300)

The MCP server allows AI agents to create posters through natural language.

// Claude Desktop configuration
{
  "mcpServers": {
    "poster-design": {
      "command": "python",
      "args": ["-m", "poster_design.mcp.server"]
    }
  }
}

To register the MCP server with Claude Code CLI:

# Basic registration (for system Python)
claude mcp add --transport stdio poster-design -- python -m poster_design.mcp.server

# For conda/virtual environments, use absolute Python path
claude mcp add --transport stdio poster-design -- /opt/anaconda3/bin/python -m poster_design.mcp.server

# Or use the console script (with full path)
claude mcp add --transport stdio poster-design -- /opt/anaconda3/bin/poster-design-mcp

Important Notes:

• If you're using conda or a virtual environment, you must use the absolute path to your Python interpreter
• To find your Python path: which python (macOS/Linux) or where python (Windows)
• Verify connection: claude mcp list - should show poster-design: ✓ Connected
• If connection fails, remove and re-register: claude mcp remove poster-design

Common Issues:

The library includes built-in background textures that can be referenced by name:

from poster_design import Canvas

canvas = Canvas(preset="instagram_post")
canvas.set_background(preset="paper_texture")
canvas.export("output.png")

from poster_design import list_preset_backgrounds

presets = list_preset_backgrounds()
print(presets)  # ['paper_texture', 'wood_grain', 'fabric_linen', 'marble', 'geometric']

You can register your own preset backgrounds at runtime:

from poster_design.core.backgrounds import register_preset_background

register_preset_background("my_texture", "/path/to/texture.jpg")
canvas.set_background(preset="my_texture")

text = TextElement(
    text="Hello World",
    font_size=48,
    font_family="Arial",
    color="#000000",
    bold=True,
    italic=False,
    align="center"
)
canvas.add(text, position="center")

image = ImageElement(
    source="path/to/image.jpg",
    size=(300, 200),
    fit="cover"  # cover, contain, or fill
)
canvas.add(image, position="center")

circle = ShapeElement(
    shape_type="circle",
    size=(100, 100),
    fill_color="#FF0000",
    border_color="#000000",
    border_width=2
)
canvas.add(circle, position=(100, 100))

from poster_design.core.layout import LayoutManager, GridLayout

# Align elements
LayoutManager.align_elements(elements, "center", canvas.size)

# Distribute elements
LayoutManager.distribute_elements(elements, "horizontal", canvas.size, spacing=50)

# Grid layout
grid = GridLayout(rows=2, cols=3, padding=20, spacing=10)

Elements support semantic position aliases:

• "center" - Centered horizontally and vertically
• "top-center", "bottom-center"
• "left-center", "right-center"
• "top-left", "top-right", "bottom-left", "bottom-right"

Or use explicit coordinates: position=(x, y) or with offsets: position="center", offset_x=50, offset_y=-30

# PNG with high DPI
canvas.export("output.png", dpi=300)

# JPG with quality
canvas.export("output.jpg", format="jpg", quality=85)

# PDF export
canvas.export("output.pdf", format="pdf")

When using MCP integration, the following tools are available:

• create_canvas - Create a new canvas
• set_background - Set canvas background (color/gradient/image)
• add_text - Add text element
• add_image - Add image element
• add_shape - Add shape element
• remove_element - Remove an element
• align_elements - Align elements
• distribute_elements - Distribute elements evenly
• export_poster - Export to file
• get_canvas_info - Get canvas information
• list_canvases - List all canvases in session
• delete_canvas - Delete a canvas

# Run all tests
pytest

# Run with coverage
pytest --cov=poster_design --cov-report=html

# Run specific test file
pytest tests/unit/test_canvas.py

# Format code
black .

# Lint
ruff .

# Type check
mypy src/

poster-design/
├── src/poster_design/
│   ├── core/           # Core SDK (no MCP dependencies)
│   │   ├── canvas.py
│   │   ├── elements.py
│   │   ├── layout.py
│   │   ├── styles.py
│   │   └── filters.py
│   ├── mcp/            # Optional MCP integration
│   │   ├── server.py
│   │   ├── handlers.py
│   │   └── session.py
│   └── utils/
│       ├── colors.py
│       └── validators.py
├── tests/
│   ├── unit/
│   └── integration/
└── examples/

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Write tests for your changes (TDD)
4. Ensure all tests pass (pytest)
5. Format your code (black .)
6. Commit your changes (git commit -m 'Add amazing feature')
7. Push to the branch (git push origin feature/amazing-feature)
8. Open a Pull Request

This project is licensed under the MIT License - see the LICENSE file for details.

• Built with Pillow for image manipulation
• MCP integration follows the Model Context Protocol
• Inspired by design tools like Figma and Canva