AI Text Generator for Inkscape

Generate and transform text in Inkscape using local AI models (Ollama, llamafile)

A powerful Inkscape extension that leverages local Large Language Models to create, modify, translate, summarize, expand, and rewrite text directly within your designs. Privacy-focused with no cloud dependencies.

📺 Demo

Click to watch the full tutorial on YouTube

---

📋 Table of Contents

• Features
• Requirements
• Installation
• Quick Start
• Usage Guide
• Configuration
• Examples
• Troubleshooting
• Contributing
• License

---

✨ Features

• 🤖 Multiple AI Providers

  • Ollama: Popular local LLM runner
  • llamafile: Single-file executable LLMs
  • Custom API: Any OpenAI-compatible endpoint

• 📝 Six Operation Modes

  Mode	Description
  Create	Generate new text from a prompt
  Modify	Change selected text based on instructions
  Translate	Convert text to another language
  Summarize	Condense text to key points
  Expand	Add detail and elaboration
  Rewrite	Improve grammar, clarity, and style

• 🎨 Rich Styling Options

  • Font family, size, weight, and style
  • Text color and decorations
  • Background boxes with customizable padding
  • Text alignment and spacing controls
  • Scale multiplier for easy resizing

• 🎯 Smart Positioning

  • 9 preset positions (center, corners, edges)
  • Cursor/selection-based placement
  • Custom X/Y offsets

• 🔧 Advanced Features

  • Auto-detect available models
  • Temperature and token limit controls
  • 8 tone presets (formal, casual, humorous, etc.)
  • Automatic markdown cleanup
  • Preserve existing text styles

---

📦 Requirements

Core Requirements

Component	Version	Purpose
Inkscape	1.0+	Vector graphics editor
Python	3.6+	Extension runtime (bundled with Inkscape)

AI Backend (Choose One)

🦙 Option 1: Ollama (Recommended)

Ollama is the easiest way to run local LLMs.

Installation:

# macOS/Linux
curl -fsSL https://ollama.com/install.sh | sh

# Windows
# Download from: https://ollama.com/download

Pull a model:

# Recommended models
ollama pull llama3.2        # Fast, good quality (2GB)
ollama pull mistral         # Excellent for text (4GB)
ollama pull gemma2          # Google's model (5GB)
ollama pull qwen2.5         # Multilingual (4GB)

Start the server:

ollama serve
# Server runs at http://localhost:11434

📦 Option 2: llamafile

llamafile packages models as single executables.

Installation:

1. Download a llamafile from Mozilla's collection
2. Make it executable and run:

# Linux/macOS
chmod +x mistral-7b-instruct-v0.2.Q4_0.llamafile
./mistral-7b-instruct-v0.2.Q4_0.llamafile --server

# Windows
# Just double-click the .llamafile or run:
mistral-7b-instruct-v0.2.Q4_0.llamafile.exe --server

Server runs at http://localhost:8080 by default.

⚙️ Option 3: Custom API

Any OpenAI-compatible API endpoint works:

• LM Studio
• LocalAI
• Text Generation WebUI (with --api flag)
• vLLM
• Any other OpenAI-compatible server

---

🚀 Installation

Step 1: Locate Your Inkscape Extensions Directory

Windows:

C:\Users\[YourUsername]\AppData\Roaming\inkscape\extensions\

macOS:

~/Library/Application Support/org.inkscape.Inkscape/config/inkscape/extensions/

Linux:

~/.config/inkscape/extensions/

💡 Tip: In Inkscape, go to Edit → Preferences → System to find your extensions directory.

Step 2: Install the Extension

1. Create the extension folder:

   mkdir -p [extensions-directory]/text_gen

2. Copy the files:

   cp text_gen.py [extensions-directory]/text_gen/
   cp text_gen.inx [extensions-directory]/text_gen/

3. Set permissions (Linux/macOS):

   chmod +x [extensions-directory]/text_gen/text_gen.py

4. Restart Inkscape

Step 3: Verify Installation

Open Inkscape and check: Extensions → Text → AI Text Generator

---

🎯 Quick Start

Generate New Text

1. Open Inkscape
2. Go to Extensions → Text → AI Text Generator
3. In the Mode tab, select Create
4. In the Prompt tab, enter your prompt:

   Write a catchy tagline for a coffee shop
   

5. Click Apply

Translate Existing Text

1. Select a text object in your document
2. Open Extensions → Text → AI Text Generator
3. Select Translate mode
4. Choose target language (e.g., "French")
5. Click Apply

---

📖 Usage Guide

Tab Overview

Tab	Purpose
Mode	Select operation (create, modify, translate, etc.)
Prompt	Enter your text generation prompt
Style	Configure font, color, and decorations
Layout	Set alignment, spacing, and positioning
Background	Add optional background box
API Config	Configure AI provider and model
Advanced	Fine-tune generation parameters

Operation Modes

🆕 Create Mode

Generate new text from scratch.

Prompt: "Write a professional bio for a graphic designer"

✏️ Modify Mode

Change selected text based on instructions.

Selected text: "Our product is good"
Prompt: "Make it more enthusiastic and professional"
Result: "Our product delivers exceptional results that exceed expectations"

🌍 Translate Mode

Convert text to another language while preserving meaning.

Selected text: "Welcome to our store"
Target language: Spanish
Result: "Bienvenido a nuestra tienda"

📝 Summarize Mode

Condense long text to key points.

📚 Expand Mode

Add detail and elaboration to short text.

🔄 Rewrite Mode

Improve grammar, clarity, and style.

Style Options

Option	Description	Default
Font Family	Any installed font	Arial
Font Size	Size in pixels	24
Font Weight	normal, bold	normal
Font Style	normal, italic	normal
Text Color	Hex color code	#000000
Text Decoration	underline, overline, line-through	none
Text Scale	Multiplier (0.5 = half, 2.0 = double)	1.0

Position Modes

┌─────────────────────────────────────┐
│ top_left    top_center    top_right │
│                                     │
│ middle_left   center   middle_right │
│                                     │
│ bottom_left bottom_center bottom_right │
└─────────────────────────────────────┘

• cursor: Place at selected object's center
• Custom offsets: Add X/Y offset to any position

Tone Presets

Tone	Effect
formal	Professional, business language
casual	Conversational, relaxed
professional	Corporate, polished
friendly	Warm, approachable
enthusiastic	Energetic, excited
humorous	Witty, playful
serious	Straightforward, no-nonsense
poetic	Artistic, metaphorical

---

⚙️ Configuration

API Configuration Tab

Setting	Description	Default
Provider	ollama, llamafile, or custom	ollama
API URL	Server address	http://localhost:11434
Model	Model name (leave empty for auto-detect)	(auto)
Auto-detect	Automatically find available model	✓

Advanced Tab

Setting	Description	Default
Temperature	Creativity (0.0 = focused, 1.0 = creative)	0.7
Max Tokens	Maximum response length	500
Remove Asterisks	Strip markdown bold/italic markers	✓
Remove Quotes	Strip surrounding quotation marks	✓
Capitalize First	Ensure first letter is uppercase	✗
Preserve Style	Keep existing text styling when modifying	✓

Provider-Specific URLs

Provider	Default URL	Notes
Ollama	http://localhost:11434	Default Ollama port
llamafile	http://localhost:8080	Default llamafile port
LM Studio	http://localhost:1234	Default LM Studio port

---

💡 Examples

Example 1: Create a Headline

Settings:

• Mode: Create
• Prompt: Write a bold headline for an eco-friendly product launch
• Tone: Enthusiastic
• Font Size: 48
• Font Weight: Bold

Result:

🌱 Revolutionize Your World with Sustainable Innovation!

Example 2: Translate for Multilingual Poster

Settings:

• Mode: Translate
• Target Language: Japanese
• Preserve Style: ✓

Original: "Innovation starts here"
Result: "イノベーションはここから始まる"

Example 3: Professional Bio

Settings:

• Mode: Create
• Prompt: Write a 2-sentence professional bio for a UX designer named Sarah
• Tone: Professional
• Max Tokens: 100

Result:

Sarah is an award-winning UX designer with over 8 years of experience crafting intuitive digital experiences. Her work spans Fortune 500 companies and innovative startups, where she transforms complex challenges into elegant, user-centered solutions.

Example 4: Expand a Tagline

Settings:

• Mode: Expand
• Tone: Friendly

Original: "Fresh coffee daily"
Result: "Every morning, we roast our carefully sourced beans to perfection, ensuring each cup delivers the rich, aromatic experience you deserve. Our commitment to freshness means your coffee is never more than 24 hours from roasting to your cup."

---

🐛 Troubleshooting

Common Issues

Extension not appearing in menu

Solutions:

1. Verify files are in the correct location:

   ls [extensions-directory]/text_gen/
   # Should show: text_gen.py, text_gen.inx

2. Check file permissions (Linux/macOS):

   chmod +x text_gen.py

3. View Inkscape error log:
   • Edit → Preferences → System → Open Error Log
4. Restart Inkscape completely

Cannot connect to Ollama

Error: Cannot connect to Ollama at http://localhost:11434

Solutions:

1. Ensure Ollama is running:

   ollama serve

2. Check if a model is installed:

   ollama list

3. Pull a model if needed:

   ollama pull llama3.2

4. Verify the URL in API Config tab

Model not found

Error: model 'xyz' not found

Solutions:

1. List available models:

   ollama list

2. Pull the required model:

   ollama pull [model-name]

3. Enable "Auto-detect model" in API Config
4. Manually enter the exact model name

Text not appearing in document

Solutions:

1. Check the position mode settings
2. Zoom out to see if text is off-canvas
3. Try "center" position mode
4. Check that the AI generated a response (look for errors)

Slow response times

Solutions:

1. Use a smaller model:

   ollama pull phi3      # 1.5GB, very fast
   ollama pull tinyllama # 600MB, fastest

2. Reduce max tokens (150-300 for simple tasks)
3. Ensure your system has enough RAM
4. Close other resource-intensive applications

Selected text not detected

Solutions:

1. Ensure you've selected a text object (not a group or path)
2. Click directly on the text with the Selection tool
3. For grouped text, enter the group first (double-click)
4. Check that the text element has actual text content

Debug Mode

Enable debug output to diagnose issues:

1. Run Inkscape from terminal:

   inkscape

2. Debug messages will appear in the terminal
3. Check the Inkscape error log: Edit → Preferences → System → Open Error Log

---

📊 Model Recommendations

Use Case	Recommended Model	Size	Speed
Quick drafts	phi3 or tinyllama	1-2GB	⚡⚡⚡
General text	llama3.2 or mistral	4GB	⚡⚡
Translation	qwen2.5	4GB	⚡⚡
Creative writing	llama3.1:8b	8GB	⚡
Best quality	llama3.1:70b	40GB	🐢

---

📁 File Structure

text_gen/
├── text_gen.py          # Main extension code
├── text_gen.inx         # Inkscape extension definition
└── README.md            # This file

---

🤝 Contributing

Contributions are welcome! Please:

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

Development Setup:

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

---

📄 License

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

---

📧 Support

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

---

🙏 Acknowledgments

• Built on Inkscape Extension API
• Powered by Ollama and llamafile
• Inspired by the need for local, privacy-respecting AI tools

---

🔄 Changelog

v1.0.0 (2024)

• ✨ Initial release
• ✅ Ollama integration
• ✅ llamafile support
• ✅ 6 operation modes
• ✅ Auto model detection
• ✅ Rich styling options
• ✅ 8 tone presets
• ✅ Background box support
• ✅ Markdown cleanup