auto-snap-mcp
Documentation
Auto-Snap MCP 📸
Automated screenshot capture and document processing for MCP Clients
Turn your screenshots into PDFs automatically! Auto-Snap lets your MCP client capture windows, process documents, and create PDFs with simple natural language commands.
🚀 Quick Start
Which Option Should I Choose?
- 🐳 Docker → Zero setup, works everywhere (recommended)
- 🪟 WSL2 → Windows users with WSL2 installed
- 🐧 Linux → Native Linux
Setup Steps
1. Get Auto-Snap:
git clone https://github.com/your-repo/auto-snap-mcp
cd auto-snap-mcp2. Install dependencies (Linux only):
# Skip this step if using Docker
sudo apt install -y wmctrl xdotool tesseract-ocr
uv sync3. Configure Claude Desktop:
Edit `~/.claude/claude_desktop_config.json` and add one of these:
🐳 Docker (Recommended - No setup needed!)
{
"mcpServers": {
"auto-snap-mcp": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "DISPLAY=:0",
"-v", "/tmp/.X11-unix:/tmp/.X11-unix:rw",
"-v", "${HOME}/auto-snap-captures:/app/captures:rw",
"mcp/auto-snap-mcp:latest"
]
}
}
}🪟 WSL2 (Windows with WSL2)
{
"mcpServers": {
"auto-snap-mcp": {
"command": "wsl.exe",
"args": [
"bash", "-c",
"cd /home/YOUR_WSL_USERNAME/auto-snap-mcp && DISPLAY=:0 /home/YOUR_WSL_USERNAME/.local/bin/uv run python server.py"
]
}
}
}🐧 Linux Native
{
"mcpServers": {
"auto-snap-mcp": {
"command": "uv",
"args": ["run", "python", "server.py"],
"cwd": "/home/YOUR_USERNAME/auto-snap-mcp",
"env": {"DISPLAY": ":0"}
}
}
}4. Restart Claude Desktop and verify it works:
- *"List all my open windows"* ← Start with this to test the connection
- *"Capture this PDF and convert to images"*
- *"Take 5 screenshots and make them into a PDF"*
✅ Working? You should see your windows listed. If not, check the 🚨 Not Working? section below.
📁 Where Your Files Are Saved
By default, all your captures go to:
~/auto-snap-captures/Full path examples:
- Linux: `/home/username/auto-snap-captures/`
- WSL2: `/home/username/auto-snap-captures/`
- Windows (Docker): `C:\Users\username\auto-snap-captures\`
✨ What It Does & Commands to Try
📋 Document Capture
- Screenshot any window or the entire screen
- Capture multi-page documents automatically
- Works with PDFs, presentations, web pages
🔍 Image Processing
- Extract text from screenshots (OCR)
- Enhance image quality automatically
- Process multiple images at once
📄 PDF Creation
- Convert screenshots to PDF instantly
- Organize files with smart naming
- Compress PDFs for smaller size
🎯 Commands to Try
"Capture this document as PDF" → Takes screenshots and creates a PDF
"Extract text from these images" → Runs OCR on screenshots
"Archive this presentation" → Screenshots all slides into one PDF
"List all my open windows" → See what you can capture
"Take 5 screenshots and make them into a PDF" → Multi-page capture
💡 Pro Tips:
- Be specific: *"Capture the Chrome window"* works better than *"take a screenshot"*
- Multi-step: *"Screenshot this presentation and extract all the text"*
- Batch work: *"Process all images in my Downloads and make PDFs"*
Want to change where files are saved? See the 🎨 Customize Your Captures section below.
🛠️ System Support
- ✅ Linux (native X11)
- ✅ WSL2 (Windows apps from Linux)
- ⚠️ macOS (with XQuartz - experimental)
🚨 Not Working?
Common fixes:
# Check dependencies
uv run python -c "from capture import check_dependencies; print(check_dependencies())"
# Fix X11 display
export DISPLAY=:0
# Test the server
uv run python server.pyStill stuck?
1. Make sure the config path is correct: `~/.claude/claude_desktop_config.json`
2. Restart Claude Desktop after config changes
3. Check Claude Desktop logs for errors
🎨 Customize Your Captures
Set where files go:
export AUTO_SNAP_OUTPUT_DIR="$HOME/Documents/Screenshots"Organize by date:
export AUTO_SNAP_USE_DATE_SUBDIRS=true
export AUTO_SNAP_INCLUDE_TIMESTAMP=trueCustom file names:
export AUTO_SNAP_FILE_NAME_TEMPLATE="doc_{page:04d}"🌍 Multiple Languages
Add more OCR languages:
sudo apt install tesseract-ocr-spa # Spanish
sudo apt install tesseract-ocr-fra # French
sudo apt install tesseract-ocr-deu # GermanThen tell Claude: *"Process this document in Spanish"*
🔐 Privacy
- Everything runs locally on your machine
- No cloud services or uploads
- Only captures what you ask for
- Automatic cleanup of temp files
Ready to automate your screenshots? Install Auto-Snap and start talking to your MCP Clients about your documents! 🚀
Frequently asked questions
What is auto-snap-mcp?
auto-snap-mcp is a Model Context Protocol (MCP) server listed in the TrackMCP directory.
How do I install auto-snap-mcp?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is auto-snap-mcp open source?
Yes — it is hosted on GitHub at https://github.com/povedaaqui/auto-snap-mcp.
Related MCP tools
Cognee is the open-source AI memory platform for agents. Give your AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
Python SQL Parser and Transpiler
MCP server that interacts with Obsidian via the Obsidian rest API community plugin
Open-source meeting transcription API for Google Meet, Microsoft Teams & Zoom. Auto-join bots, real-time WebSocket transcripts, MCP server for AI agents. Self-host or use hosted SaaS.
A super light-weight embedded code search engine CLI (AST based) that just works - improves speed and efficiency for coding agent 🌟 Star if you like it!
Official MiniMax Model Context Protocol (MCP) server that enables interaction with powerful Text to Speech, image generation and video generation APIs.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP