loading…
Search for a command to run...
loading…
▸ TL;DR
Use the official SDK in your language (TypeScript, Python, Go), define tools with JSON Schema, expose them via stdio or HTTP. Submit to unyly.org/submit so other users can discover it.
Read MCP explainerThe MCP spec defines a JSON-RPC dialect over stdio or HTTP/SSE. Official SDKs exist for TypeScript, Python and Go (github.com/modelcontextprotocol). Pattern: 1) implement the Server interface, 2) register tools with name, description, input schema, 3) handle tool calls by returning results. For TypeScript: `npm install @modelcontextprotocol/sdk` and use the StdioServerTransport. Test locally with Claude Desktop by pointing claude_desktop_config.json at your binary. When ready, submit to unyly.org/submit — the catalog auto-verifies and lists your MCP.
Partially. MCP solves "tool plumbing" — connecting AI to tools with one protocol. LangChain is a broader agent framework with chains, memory, RAG. Use MCP if you want simple, language-agnostic tool integration. Use LangChain if you need a full agent runtime in Python/JS.
API keys are passed to the MCP server as environment variables when it starts. The MCP server reads them and uses them to call upstream APIs. The LLM never sees the raw credentials — only sees the tools they enable.