Install LLM-MCP STDIO Integration
The Leedz MCP (Model Context Protocol) plugin connects AI chat clients
like Claude Desktop and LM Studio directly to your local Leedz database.
This lets you manage clients and bookings using natural language in
your favorite AI chat interface.
Example: "show all the April birthday parties I've ever booked"
THIS IS A STDIO PLUGIN
- Runs locally on your computer as a background process
- Communicates via stdin/stdout (text streams)
- Your data NEVER leaves your machine
- Near-instant response times (no network latency)
- No monthly hosting fees or cloud infrastructure
LIMITATION:
Not all AI LLM clients support STDIO plugins. Currently tested on:
- Anthropic Claude Desktop
- LM Studio
BEFORE INSTALLING:
1. Node.js must be installed on your system
- Required ONLY for the MCP plugin (runs as .js script)
- The Leedz server.exe INCLUDES Node runtime (no separate install needed)
- Check if Node is installed: Open command prompt, type "node --version"
- If not installed: Download from https://nodejs.org
2. The Leedz Server must be installed and running
- See server/INSTALL_INSTRUCTIONS.txt
- Verify server is running on http://localhost:3000
3. An AI chat client that supports MCP (Claude Desktop or LM Studio)
INSTALL IN CLAUDE DESKTOP (Windows)
STEP 1: Locate your Claude Desktop config file
The config file is at:
C:\Users\\AppData\Roaming\Claude\claude_desktop_config.json
Replace
with your Windows username.
STEP 2: Open the config file in a text editor
- Right-click the file, select "Open with > Notepad"
- If the file doesn't exist, create it
STEP 3: Add the Leedz MCP server definition
If your config file is empty, paste this (UPDATE THE PATHS):
{
"mcpServers": {
"leedz-mcp": {
"command": "C:\\Program Files\\nodejs\\node.exe",
"args": ["C:\\Users\\\\Desktop\\WKG\\INVOICER\\server\\mcp\\mcp_server.js"]
}
},
"globalShortcut": ""
}
If your config file already has content, add the "leedz-mcp" section
inside the existing "mcpServers" object:
{
"mcpServers": {
"leedz-mcp": {
"command": "C:\\Program Files\\nodejs\\node.exe",
"args": ["C:\\Users\\*YourUsername*\\Desktop\\WKG\\INVOICER\\server\\mcp\\mcp_server.js"]
},
"other-mcp-server": {
...existing servers...
}
},
"globalShortcut": ""
}
PATH NOTES:
- Replace with your actual Windows username
- Use DOUBLE BACKSLASHES (\\) in Windows paths
- Update the path to match where you installed the Leedz server
- Common install locations:
C:\\Program Files\\Leedz\\mcp\\mcp_server.js
C:\\Users\*YourUsername*\Desktop\\Leedz\\server\\mcp\\mcp_server.js
STEP 4: Save the file and restart Claude Desktop
STEP 5: Verify installation
Open Claude Desktop and type: "list all clients from the Leedz"
If configured correctly, you'll see your client list.
INSTALL IN LM STUDIO:
STEP 1: Open LM Studio
STEP 2: Navigate to the Integrations panel
- Look for the right-side panel labeled "Integrations" or "Program"
- If not visible, check the View menu or toolbar
STEP 3: Add MCP Server
- Click the dropdown menu: "Install"
- Select "MCP"
- Click "Edit mcp.json"
STEP 4: Add the Leedz MCP server definition
Paste this into the editor (UPDATE THE PATH):
{
"mcpServers": {
"leedz-mcp": {
"command": "C:\\Program Files\\nodejs\\node.exe",
"args": [
"C:\\Users\\*YourUsername*\\Desktop\\WKG\\INVOICER\\server\\mcp\\mcp_server.js"
]
}
}
}
If you already have other MCP servers configured, add "leedz-mcp"
to the existing "mcpServers" object.
PATH NOTES:
- Replace with your actual Windows username
- Use DOUBLE BACKSLASHES (\\) in Windows paths
- Update the path to match your Leedz server installation
STEP 5: Save the file
STEP 6: Restart LM Studio
STEP 7: Verify installation
Open a chat and type:
"show me all bookings from the Leedz"
If configured correctly, you'll see your booking list.
SETTINGS:
For detailed API endpoint documentation see:
server/README.md
For MCP server configuration:
server/mcp/mcp_server_config.json
For logs and debugging:
server/mcp/mcp_server.log
CONFIG file location:
server/mcp/mcp_server_config.json
- database.apiUrl: URL of Leedz server (default: http://localhost:3000)
- llm.api-key: Anthropic API key for natural language translation
- llm.provider: Claude model used for translation (default: claude-opus-4-1)
- logging.file: Path to log file (default: ./mcp_server.log)
If your Leedz server runs on a different port:
- Edit server/mcp/mcp_server_config.json
- Change "apiUrl": "http://localhost:3000" to your port
- Restart the AI client
LOG FILE:
All plugin activity is logged to server/mcp/mcp_server.log
Check this file for debugging and troubleshooting.