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¶
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¶
- Completely restart your application
- Ask:
"Can you call session_information() to show my napari session?"
- 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¶
- Claude Desktop - Most popular choice for general use
- Claude Code - CLI integration for development workflows
💻 IDE Integrations¶
- Cursor IDE - AI-powered coding with napari
- Cline - VS Code and Cursor extensions
🌐 Other Platforms¶
- Gemini CLI & Codex - Google Gemini and OpenAI Codex setup
- Python Integration - Custom scripts for workflow automation
- ChatGPT - Why ChatGPT doesn't work and what to use instead
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
Configuration not detected
Permission errors
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¶
- Choose your platform based on your primary use case
- Follow the specific setup guide for detailed instructions
- Test the integration with our provided examples
- Explore advanced features once basic setup works
Ready to connect your AI assistant? Choose your platform above and let's get started! 🚀