• Features
• Requirements
• Installation
• Quick Start
• Usage Guide
• Configuration
• Provider Setup
• Customization
• Troubleshooting
• Contributing
• License

• 🤖 Multiple AI Providers

  • OpenAI: GPT-4, GPT-4 Turbo, GPT-3.5
  • Anthropic: Claude 3.5 Sonnet, Claude 3 Opus
  • Google: Gemini 1.5 Flash, Gemini 1.5 Pro
  • Ollama: Local models (Llama, Mistral, etc.)

• 🎨 Rich Style Options

  • Style presets (minimal, detailed, flat, outline, geometric, etc.)
  • Color schemes (monochrome, warm, cool, pastel, vibrant, etc.)
  • Complexity levels (simple, medium, complex)
  • Stroke style control

• 📐 Flexible Sizing

  • Preset sizes (small, medium, large, xlarge)
  • Custom dimensions
  • Aspect ratios (square, landscape, portrait, widescreen, banner)

• 🎯 Smart Positioning

  • Center, origin, or next to selection
  • Multiple variations side-by-side
  • Automatic grouping with custom names

• ⚙️ Advanced Features

  • Selection context awareness
  • Prompt presets (icon, logo, diagram, pattern, etc.)
  • Gradient and animation support
  • Accessibility options (title/desc)
  • Prompt history tracking
  • Retry logic for reliability

💡 Tip: In Inkscape: Edit → Preferences → System shows extensions path

1. Create extension folder:

   mkdir -p [extensions-directory]/svg_maker

2. Copy files:

   cp svg_llm.py [extensions-directory]/svg_maker/
   cp svg_llm.inx [extensions-directory]/svg_maker/

3. Restart Inkscape

1. Open Inkscape
2. Go to Extensions → Generate → SVG LLM Generator
3. The extension dialog should appear

1. Open Inkscape
2. Go to Extensions → Generate → SVG LLM Generator
3. Select a provider (e.g., OpenAI)
4. Enter your API key
5. Type a prompt: A simple house icon with a chimney
6. Click Apply

Result: An AI-generated SVG house icon appears on your canvas!

Three ways to provide API keys (in priority order):

Enter key directly in the Provider tab each time.

1. Enter API key in Provider tab
2. Check "Save API key"
3. Key is stored in .config.json

Edit .config.json in the extension directory:

{
 "api_keys": {
 "openai": "sk-your-openai-key",
 "anthropic": "sk-ant-your-anthropic-key",
 "google": "your-google-api-key"
 },
 "last_provider": "openai"
}

The extension automatically saves your prompts to .svg_llm_history.json:

[
 {
 "prompt": "A simple house icon",
 "timestamp": "2024年01月15日T10:30:00",
 "width": 400,
 "height": 400,
 "provider": "openai",
 "model": "gpt-4-turbo",
 "style": "minimal",
 "color_scheme": "any"
 }
]

1. Get API key from platform.openai.com/api-keys
2. Select provider: openai
3. Recommended models: gpt-4-turbo, gpt-4o, gpt-3.5-turbo

1. Get API key from console.anthropic.com
2. Select provider: anthropic
3. Recommended models: claude-3-5-sonnet-20241022, claude-3-opus-20240229

1. Get API key from aistudio.google.com/apikey
2. Select provider: google
3. Recommended models: gemini-1.5-flash, gemini-1.5-pro

1. Install Ollama from ollama.com
2. Pull a model:

   ollama pull llama3.1

3. Start Ollama:

   ollama serve

4. Select provider: ollama
5. No API key needed!

Edit the class constants in svg_llm.py:

class SVGLLMGenerator(inkex.EffectExtension):
 CONFIG_FILENAME = '.config.json' # Config file name
 HISTORY_FILENAME = '.svg_llm_history.json' # History file name
 MAX_HISTORY = 50 # Max history entries

Add new presets in the build_prompt method:

preset_prompts = {
 'icon': "Create a simple, recognizable icon...",
 'illustration': "Create an artistic illustration...",
 # Add your custom preset:
 'scientific': "Create a scientific diagram with labels and annotations...",
}

Modify style descriptions in build_prompt:

style_descriptions = {
 'minimal': 'Use a minimal, clean design...',
 # Add custom style:
 'blueprint': 'Use technical blueprint style with white lines on blue background...',
}

Add color schemes in build_prompt:

color_descriptions = {
 'monochrome': 'Use only one color...',
 # Add custom scheme:
 'synthwave': 'Use synthwave/retrowave colors (pink, cyan, purple on dark)...',
}

Edit the system message in API call methods:

def call_openai_api(self, prompt, api_key):
 # ...
 data = {
 'messages': [
 {
 'role': 'system',
 'content': 'Your custom system prompt here...'
 },
 # ...
 ]
 }

1. Check error log: Edit → Preferences → System → Open Error Log
2. Test API independently: Use curl or Postman to verify API access
3. Start simple: Test with basic prompts before complex ones
4. Check history: Review .svg_llm_history.json for past successful prompts

svg_maker/
├── svg_llm.py # Main extension code
├── svg_llm.inx # Inkscape extension definition
├── README.md # This file
├── LICENSE # MIT License
├── .config.json # Saved API keys (auto-created)
└── .svg_llm_history.json # Prompt history (auto-created)

1. Be specific: "A red apple with a leaf" > "an apple"
2. Mention style: "flat design icon of..." or "detailed illustration of..."
3. Specify colors: "using blue and orange colors"
4. Define complexity: "simple/minimal" or "detailed/intricate"

1. Use presets: They include optimized instructions
2. Start with medium complexity: Adjust based on results
3. Lower temperature for consistency: 0.3-0.5 for predictable results
4. Higher temperature for creativity: 0.8-1.0 for variety

1. GPT-4 models produce best SVG: More reliable than GPT-3.5
2. Claude excels at complex diagrams: Good for flowcharts
3. Gemini is fast and affordable: Good for simple icons
4. Local Ollama varies by model: llama3.1 recommended

Contributions are welcome!

1. Fork the repository
2. Create a feature branch (git checkout -b feature/new-provider)
3. Commit changes (git commit -m 'Add new AI provider')
4. Push to branch (git push origin feature/new-provider)
5. Open a Pull Request

Development Setup:

git clone https://github.com/YouvenZ/svg_maker_ink.git
cd svg_maker_ink
# Symlink for testing
ln -s $(pwd) ~/.config/inkscape/extensions/svg_maker

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

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: youvenz.pro@gmail.com

• Built on Inkscape Extension API
• Powered by OpenAI, Anthropic, Google, and Ollama
• Inspired by the need for AI-assisted vector graphics creation

• ✨ Initial release
• ✅ OpenAI, Anthropic, Google, Ollama support
• ✅ Multiple style and color options
• ✅ Flexible sizing with aspect ratios
• ✅ Selection context awareness
• ✅ Multiple variations generation
• ✅ Prompt history tracking
• ✅ API key persistence
• ✅ Retry logic for reliability