Skip to content

LLM Application Integrations

Connect napari MCP server with your favorite AI assistant or development environment using our automated CLI installer.

🚀 Quick Installation

All integrations use the same simple command:

# Install napari-mcp
pip install napari-mcp

# Auto-configure your application
napari-mcp-install <application-name>

Supported Platforms

Platform Command Status Guide
Claude Desktop napari-mcp-install claude-desktop ✅ Full Support Setup →
Claude Code napari-mcp-install claude-code ✅ Full Support Setup →
Cursor IDE napari-mcp-install cursor ✅ Full Support Setup →
Cline napari-mcp-install cline-vscode or cline-cursor ✅ Full Support Setup →
Gemini / Codex napari-mcp-install gemini or codex ✅ Full Support Setup →
Python Custom script ✅ Full Support Guide →
ChatGPT N/A ❌ Not Supported Why? →

Feature Comparison

Feature Claude Desktop Claude Code Cursor Cline Gemini/Codex
Visual napari window ✅ Full ✅ Full ✅ Full ✅ Full ✅ Full
All MCP tools ✅ 20+ tools ✅ 20+ tools ✅ 20+ tools ✅ 20+ tools ✅ 20+ tools
File system access ✅ Full ✅ Full ✅ Full ✅ Full ✅ Full
Code execution ✅ Yes ✅ Yes ✅ Yes ✅ Yes ✅ Yes
Package installation ✅ Yes ✅ Yes ✅ Yes ✅ Yes ✅ Yes
Setup complexity 🟢 Easy 🟢 Easy 🟢 Easy 🟢 Easy 🟢 Easy
Best for General use Development AI coding VS Code users Advanced users

Installation Workflow

Step 1: Install Package

pip install napari-mcp

Step 2: Auto-Configure

# For your specific application
napari-mcp-install <app-name>

# See all options
napari-mcp-install --help

# Preview changes before applying
napari-mcp-install <app-name> --dry-run

Step 3: Restart & Test

  1. Completely restart your application
  2. Ask: "Can you call session_information() to show my napari session?"
  3. Success! ✅

Installation Options

The CLI installer supports several options:

# Use your Python environment instead of uv
napari-mcp-install claude-desktop --persistent

# Custom Python path
napari-mcp-install claude-desktop --python-path /path/to/python

# Preview changes only
napari-mcp-install claude-desktop --dry-run

# Force update without prompts
napari-mcp-install claude-desktop --force

# Install for all applications
napari-mcp-install all

Management Commands

# List all installations
napari-mcp-install list

# Uninstall from an application
napari-mcp-install uninstall claude-desktop

# Uninstall from all
napari-mcp-install uninstall all

Platform-Specific Guides

Choose your application for detailed setup instructions:

🖥️ Desktop Applications

💻 IDE Integrations

  • Cursor IDE - AI-powered coding with napari
  • Cline - VS Code and Cursor extensions

🌐 Other Platforms

Common Configuration

All platforms support environment variables for advanced configuration:

export QT_QPA_PLATFORM=offscreen  # For headless servers
export NAPARI_ASYNC=1             # Enable async operations
export MCP_LOG_LEVEL=INFO         # Debug MCP communication

Troubleshooting

CLI Installer Issues

napari-mcp-install: command not found

# Reinstall package
pip install --force-reinstall napari-mcp

# Verify
napari-mcp-install --version

Configuration not detected

# List what would be configured
napari-mcp-install <app> --dry-run

# Check current installations
napari-mcp-install list

Permission errors

# Check config locations
napari-mcp-install list

# Fix permissions (macOS/Linux)
chmod 644 <config-file-path>

Application-Specific Issues

Each integration guide has troubleshooting sections for platform-specific problems:

  • Claude Desktop - Config file location, restart issues
  • Claude Code - CLI configuration, environment setup
  • Cursor - Project vs global installation
  • Cline - Extension detection, VS Code variants
  • Gemini/Codex - TOML config, API setup

→ See Troubleshooting Guide for comprehensive help

Advanced: Python Scripting

For batch processing or custom pipelines, see Python Scripts to use napari MCP with any LLM in your own code.

Next Steps

  1. Choose your platform based on your primary use case
  2. Follow the specific setup guide for detailed instructions
  3. Test the integration with our provided examples
  4. Explore advanced features once basic setup works

Ready to connect your AI assistant? Choose your platform above and let's get started! 🚀