MCP Client Setup Overview
SkyFi MCP Server integrates seamlessly with multiple AI development environments. Choose your preferred client and follow the setup guide to get started.
Supported AI Clients
Claude Desktop
Anthropic's official desktop application with full MCP support.
Best for: General use, AI conversations, content creation Features: Native MCP integration, intuitive interface, regular updates Setup Time: ~5 minutes
Cursor AI
AI-powered code editor with enhanced development features.
Best for: Software development, code analysis, programming tasks Features: Code completion, debugging, refactoring with satellite context Setup Time: ~5 minutes
Windsurf
Advanced AI coding assistant with collaborative features.
Best for: Team development, collaborative coding, enterprise use Features: Multi-user support, advanced code intelligence, enterprise security Setup Time: ~10 minutes
VSCode
Popular code editor with MCP extension support.
Best for: Developers, extensible workflows, custom integrations Features: Rich extension ecosystem, customizable interface, debugging tools Setup Time: ~10 minutes
Quick Comparison
| Feature | Claude Desktop | Cursor AI | Windsurf | VSCode |
|---|---|---|---|---|
| Native MCP | Built-in | Extension | Extension | Extension |
| Code Editing | Limited | Advanced | Advanced | Advanced |
| AI Chat | Primary | Integrated | Integrated | Extension |
| Debugging | None | Full | Full | Full |
| Team Features | Personal | Limited | Full | Extensions |
| Deployment | Simple | Simple | Complex | Simple |
Configuration Methods
1. Global Configuration
Install and configure SkyFi MCP system-wide for all projects.
Pros:
- Works across all projects
- Single setup process
- Consistent behavior
Cons:
- Version conflicts possible
- Hard to customize per project
2. Project-Specific Configuration
Configure SkyFi MCP for individual projects.
Pros:
- Project isolation
- Custom configurations
- Version control friendly
Cons:
- Setup per project
- Potential duplication
3. Hybrid Approach (Recommended)
Global installation with project-specific configurations.
Best of both worlds: System-wide availability with project customization.
Prerequisites
Before setting up any MCP client, ensure you have:
Required
- SkyFi MCP Server installed (Installation Guide)
- SkyFi API credentials from your account
- Python 3.11+ with virtual environment
- Your chosen AI client downloaded and installed
Optional
- OpenWeatherMap API key for weather integration
- Docker for containerized deployment
- Git for version control integration
Common Configuration Elements
All MCP clients require similar configuration elements:
Command Configuration
{
"command": "/path/to/python",
"args": ["-m", "mcp_skyfi"],
"cwd": "/path/to/project"
}
Environment Variables
{
"env": {
"SKYFI_API_KEY": "your-email@example.com:your-api-key-hash",
"SKYFI_URL": "https://app.skyfi.com/platform-api/pricing",
"OPENWEATHER_API_KEY": "optional-weather-key"
}
}
Transport Options
- STDIO (default): Direct process communication
- HTTP: Web server mode for browser-based clients
- SSE: Server-sent events for real-time updates
Setup Process Overview
Each client setup follows this general pattern:
- ** Install Prerequisites** - Client software and extensions
- ** Configure Server** - MCP server connection settings
- ** Set Credentials** - API keys and authentication
- ** Test Integration** - Verify tools are available
- ** Start Using** - Run test commands and examples
Platform-Specific Notes
macOS
- Configuration files in
~/Library/Application Support/ - Use
which pythonto find Python path - May need to allow app permissions in System Preferences
Windows
- Configuration files in
%APPDATA%\directories - Use
where pythonto find Python path - May need to run as administrator for global configs
Linux
- Configuration files in
~/.config/or~/.local/ - Package manager installations may affect paths
- Check distribution-specific documentation
Security Considerations
API Key Management
- Store securely: Use environment variables, not hardcoded values
- Limit scope: Use minimal required permissions
- Rotate regularly: Update keys periodically
- Never commit: Don't include keys in version control
Network Security
- Use HTTPS: All API calls are encrypted
- Verify certificates: Enable SSL verification
- Monitor usage: Track API calls and usage patterns
Access Control
- User-specific configs: Avoid shared credentials
- Project isolation: Separate configs per project when needed
- Audit access: Monitor who has access to what
Troubleshooting Quick Guide
Common Issues
** "Server not found"**
- Check Python path is correct
- Verify working directory exists
- Ensure MCP server is installed
** "0 tools enabled"**
- Verify API credentials are valid
- Check environment variable format
- Restart client after configuration
** "Connection timeout"**
- Check internet connectivity
- Verify firewall settings
- Try increasing timeout values
Debug Commands
# Test server directly
python -m mcp_skyfi --config-check
# Verify tool discovery
python -c "from mcp_skyfi.servers.main import main_mcp; print(len(main_mcp.list_tools()))"
# Check environment variables
env | grep SKYFI
Performance Optimization
Client-Specific Tips
- Claude Desktop: Close other resource-intensive apps
- Cursor AI: Disable unnecessary extensions
- Windsurf: Optimize team sync settings
- VSCode: Use workspace settings for better performance
General Optimization
- Use local environment variables when possible
- Enable connection pooling in configuration
- Set appropriate timeout values
- Monitor memory usage during development
Next Steps
- Choose your client from the options above
- Follow the detailed setup guide for your chosen client
- Test the integration with our verification steps
- Explore advanced features and customization options
Need Help?
- Troubleshooting Guide (coming soon) - Detailed problem solving
- GitHub Issues - Report problems
- Community Discussions - Get help from others
Ready to set up your AI client? Choose your preferred option above and follow the detailed setup guide!