MCP Server for Transcripter

A Model Context Protocol (MCP) server implementation for the Transcripter project. This package provides tools and resources for AI-powered features using the MCP standard.

Features

Tools

• test-api: Test API endpoints and return the results
• transcription-search: Search transcriptions with filtering and pagination
• transcription-summary: Generate a summary of a transcription using AI

Resources

• transcription://{id}: Access transcription data by ID
• analysis://{id}: Access analysis data by ID

Requirements

• Node.js >= 18.0.0
• npm >= 7.0.0

Installation

npm install

Building

# Build for both ESM and CommonJS
npm run build

# Build for ESM only
npm run build:esm

# Build for CommonJS only
npm run build:cjs

Running

# Start the MCP server on the default port (3500)
npm run server

# Start the MCP server on a custom port
npm run server 4000

Testing

npm test

Usage Examples

Using the test-api tool

import { Client } from "@modelcontextprotocol/sdk/client";
import { SSEClientTransport } from "@modelcontextprotocol/sdk/client/sse";

async function testApiEndpoint() {
  // Connect to the MCP server
  const transport = new SSEClientTransport("http://localhost:3500/sse", "http://localhost:3500/message");
  const client = new Client();
  await client.connect(transport);
  
  // Use the test-api tool
  const result = await client.tools.execute("test-api", {
    endpoint: "transcriptions",
    method: "GET",
  });
  
  console.log(result);
}

Using the transcription resource

import { Client } from "@modelcontextprotocol/sdk/client";
import { SSEClientTransport } from "@modelcontextprotocol/sdk/client/sse";

async function getTranscription(id: number) {
  // Connect to the MCP server
  const transport = new SSEClientTransport("http://localhost:3500/sse", "http://localhost:3500/message");
  const client = new Client();
  await client.connect(transport);
  
  // Access the transcription resource
  const transcription = await client.resources.get(`transcription://${id}`);
  
  console.log(transcription);
}

Integration with Transcripter

This MCP server integrates with the Transcripter project to provide AI-powered features for transcriptions and analyses. It serves as a standardized interface for AI model interactions.

Project Structure

• src/cli.ts: Command-line interface for starting the MCP server
• src/tools/: Implementation of MCP tools
• src/resources/: Implementation of MCP resource providers
• src/tests/: Tests for tools and resources

License

MIT