Overview
What is my-mcp-server?
my-mcp-server is a Model Context Protocol (MCP) server built with the mcp-framework. It provides a scaffold for defining and running custom MCP tools, intended for developers building AI–agent tooling that integrates with MCP-compatible clients such as Claude Desktop.
How to use my-mcp-server?
Install dependencies with npm install, then build the project with npm run build. After building, run the server via node /absolute/path/to/dist/index.js or, after publishing to npm, with npx my-mcp-server. Configure the server in Claude Desktop by adding the appropriate JSON entry to claude_desktop_config.json.
Key features of my-mcp-server
- Built on the mcp-framework for rapid MCP tool development.
- Includes an example tool (
ExampleTool.ts) to use as a template. - CLI commands to generate new tools (e.g.,
mcp add tool my-tool). - Tool schemas defined with Zod for input validation.
- Supports local development and npm publishing for distribution.
Use cases of my-mcp-server
- Rapid prototyping of custom MCP tools for AI assistant workflows.
- Adding file‑handling, data‑processing, or API‑client tools to Claude Desktop.
- Publishing reusable MCP servers to the npm registry for team or public use.
FAQ from my-mcp-server
What runtime environment does my-mcp-server require?
Node.js is required. Dependencies are installed via npm install.
How do I add a new tool to my-mcp-server?
Use the CLI command mcp add tool <tool-name> from the project root, or manually create a tool file in src/tools/ following the example structure.
How do I test my-mcp-server locally before publishing?
Build the project with npm run build, then run npm link and execute my-mcp-server on the command line.
How do I connect my-mcp-server to Claude Desktop?
Add a JSON entry to the claude_desktop_config.json file under mcpServers. For local development use the compiled dist/index.js path; after publishing use npx my-mcp-server.
Where does my-mcp-server store data?
The README does not discuss data storage. The server itself does not include built‑in persistence; data handling depends on the tools you implement.
