Skip to main content
Glama

Local MCP Server

by yulongwanxue
QUICKSTART.md3 kB
# Quick Start Guide ## Installation (5 minutes) ### 1. Install Dependencies ```bash pip install mcp aiofiles pyyaml ``` ### 2. Test the Server ```bash # Test basic functionality python examples/test_client.py ``` You should see output showing: - 18 registered tools - Calculator operations working - File operations working - Error handling working ## Usage with LangGraph ### Basic Example ```python import asyncio import sys from pathlib import Path from mcp import ClientSession, StdioServerParameters from mcp.client.stdio import stdio_client async def main(): # Start MCP server server_params = StdioServerParameters( command=sys.executable, args=[str(Path("src/server.py"))], ) # Connect and use tools async with stdio_client(server_params) as (read, write): async with ClientSession(read, write) as session: await session.initialize() # Call calculator tool result = await session.call_tool( "calculator_add", {"a": 10, "b": 5} ) print(result.content[0].text) # Output: {"result": 15.0, "operation": "add", ...} # Call file tool result = await session.call_tool( "file_write", {"file_path": "test.txt", "content": "Hello!"} ) print(result.content[0].text) # Output: {"success": true, "bytes_written": 6, ...} asyncio.run(main()) ``` ## Available Tools ### Calculator (10 tools) - `calculator_add`, `calculator_subtract`, `calculator_multiply`, `calculator_divide` - `calculator_power`, `calculator_sqrt`, `calculator_log` - `calculator_sin`, `calculator_cos`, `calculator_tan` ### File Operations (8 tools) - `file_read`, `file_read_lines`, `file_exists`, `file_list` - `file_write`, `file_append`, `file_create`, `file_delete` ## Key Features **Security:** - All file operations sandboxed to `workspace/` directory - Only relative paths accepted - File size limits (10MB default) - Extension filtering **Configuration:** - Edit `config/config.yaml` or use environment variables - `MCP_DEBUG=true` for verbose logging - `MCP_LOG_LEVEL=DEBUG` for detailed logs **Logging:** - Logs written to `logs/mcp_server.log` - Console output to stderr (stdout reserved for MCP protocol) ## Next Steps 1. Run `python examples/langgraph_integration.py` for more examples 2. Read `README.md` for full documentation 3. Add your own tools in `src/tools/` following the pattern 4. Configure `config/config.yaml` for your needs ## Troubleshooting **Server won't start?** - Check Python 3.10+ - Install dependencies: `pip install mcp aiofiles pyyaml` **File operation fails?** - Use relative paths: `data/file.txt` not `/abs/path` - Check workspace exists: `workspace/` directory - Verify in sandbox: paths must be under `workspace/` **Tool not found?** - Restart server after adding new tools - Check tool imported in `src/tools/__init__.py`

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/yulongwanxue/local-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server