# MCP Client Configuration Guide
This guide covers configuration for various MCP clients beyond Claude Desktop. For Claude Desktop configuration, see the [Claude Desktop Setup Guide](./claude_desktop_setup.md).
## 🔷 Cursor Configuration
Cursor is a VS Code fork with AI capabilities and MCP support.
> 📖 **Complete Setup Guide**: Detailed Cursor configuration instructions available at [playbooks.com](https://playbooks.com/mcp/stevereiner-alfresco-content-services#cursor-setup)
### For Users (PyPI Installation)
If you installed the package via PyPI with pipx:
```json
{
"mcpServers": {
"python-alfresco-mcp-server": {
"command": "python-alfresco-mcp-server",
"args": ["--transport", "stdio"],
"env": {
"ALFRESCO_URL": "http://localhost:8080",
"ALFRESCO_USERNAME": "admin",
"ALFRESCO_PASSWORD": "admin"
}
}
}
}
```
### For Developers (Source Installation)
If you're using the source code with UV:
```json
{
"mcpServers": {
"python-alfresco-mcp-server": {
"command": "uv",
"args": ["run", "python-alfresco-mcp-server"],
"cwd": "/path/to/python-alfresco-mcp-server",
"env": {
"ALFRESCO_URL": "http://localhost:8080",
"ALFRESCO_USERNAME": "admin",
"ALFRESCO_PASSWORD": "admin"
}
}
}
}
```
## ⚡ Claude Code Configuration
Claude Code is Anthropic's VS Code extension with MCP support.
> 📖 **Complete Setup Guide**: Claude Code configuration instructions at [playbooks.com](https://playbooks.com/mcp/stevereiner-alfresco-content-services#claude-code-setup)
### For Users (PyPI Installation)
```bash
claude mcp add-json "python-alfresco-mcp-server" '{
"command": "python-alfresco-mcp-server",
"args": ["--transport", "stdio"],
"env": {
"ALFRESCO_URL": "http://localhost:8080",
"ALFRESCO_USERNAME": "admin",
"ALFRESCO_PASSWORD": "admin"
}
}'
```
### For Developers (Source Installation)
```bash
claude mcp add-json "python-alfresco-mcp-server" '{
"command": "uv",
"args": ["run", "python-alfresco-mcp-server"],
"cwd": "/path/to/python-alfresco-mcp-server",
"env": {
"ALFRESCO_URL": "http://localhost:8080",
"ALFRESCO_USERNAME": "admin",
"ALFRESCO_PASSWORD": "admin"
}
}'
```
## 🔧 Other MCP Clients
For any MCP-compatible client, use these connection parameters based on your installation method:
### PyPI Installation (Users)
- **Command**: `python-alfresco-mcp-server` (assumes pipx installation)
- **Args**: `["--transport", "stdio"]`
- **Transport Options**:
- STDIO (default) - Direct MCP protocol
- HTTP (add `--port 8001`) - RESTful API
- SSE (add `--port 8003`) - Server-Sent Events
### Source Installation (Developers)
- **Command**: `uv`
- **Args**: `["run", "python-alfresco-mcp-server"]`
- **Working Directory**: Path to cloned repository
- **Transport Options**: Same as above
### Traditional Python Installation
If using traditional pip in a virtual environment:
- **Command**: `/path/to/venv/bin/python-alfresco-mcp-server` (full path to executable)
- **Args**: `["--transport", "stdio"]`
- **Transport Options**: Same as above
## 🔧 Environment Variables
All clients need these environment variables configured:
| Variable | Default | Description |
|----------|---------|-------------|
| `ALFRESCO_URL` | `http://localhost:8080` | Alfresco server URL |
| `ALFRESCO_USERNAME` | `admin` | Username for authentication |
| `ALFRESCO_PASSWORD` | `admin` | Password for authentication |
| `ALFRESCO_VERIFY_SSL` | `false` | Verify SSL certificates |
| `ALFRESCO_TIMEOUT` | `30` | Request timeout (seconds) |
### Windows-Specific Variables (if needed)
For Windows systems experiencing character encoding issues:
```json
"env": {
"ALFRESCO_URL": "http://localhost:8080",
"ALFRESCO_USERNAME": "admin",
"ALFRESCO_PASSWORD": "admin",
"PYTHONIOENCODING": "utf-8",
"PYTHONLEGACYWINDOWSSTDIO": "1"
}
```
## 🚀 Transport Options
The MCP server supports three transport protocols:
### STDIO (Default)
- **Fastest** and most efficient
- Direct MCP protocol communication
- Recommended for most use cases
- Args: `["--transport", "stdio"]` (optional, it's the default)
### HTTP
- RESTful API interface
- Useful for web services and testing
- Args: `["--transport", "http", "--port", "8001"]`
### SSE (Server-Sent Events)
- Real-time streaming updates
- Good for live monitoring
- Args: `["--transport", "sse", "--port", "8003"]`
## 🧪 Testing Your Configuration
After setting up your MCP client:
1. **Start Your Client**: Launch your MCP-enabled application
2. **Check Connection**: Look for "python-alfresco-mcp-server" in connected servers
3. **Test Basic Functionality**:
- Try the `repository_info` tool to verify connection
- Run a simple `search_content` query
- Check that all 15 tools are available
## 🛠️ Troubleshooting
### Common Issues
1. **Command Not Found**
- Ensure the package is installed correctly
- For pipx: Run `pipx list` to verify installation
- For source: Ensure UV is installed and working directory is correct
2. **Connection Failures**
- Check Alfresco server is running
- Verify environment variables are set correctly
- Test connection with `curl http://localhost:8080/alfresco`
3. **Permission Errors**
- Verify Alfresco username/password
- Check that user has appropriate permissions
- Try with admin credentials first
4. **Character Encoding (Windows)**
- Add Windows-specific environment variables
- Ensure UTF-8 encoding is configured
### Getting Help
- 📚 **Documentation**: Complete guides in [`../docs/`](./README.md)
- 🛠️ **Troubleshooting**: [Troubleshooting Guide](./troubleshooting.md)
- 🐛 **Issues**: [GitHub Issues](https://github.com/stevereiner/python-alfresco-mcp-server/issues)
## ⚠️ Security Notes
- **Never commit configuration files** with real credentials to version control
- **Use environment variables** for production deployments
- **Consider using .env files** for local development (they're ignored by git)
- **Use strong passwords** for production Alfresco servers