Model Context Protocol (MCP) server for CUFinder API - providing seamless access to people search, company search, and local business search capabilities.
This NestJS-based MCP server exposes CUFinder's powerful search APIs through the Model Context Protocol, allowing AI assistants and other MCP clients to search for people, companies, and local businesses.
Search for people using various filters including:
- Personal information (name, location)
- Job details (title, role, level)
- Company information (name, industry, size, revenue)
Search for companies with filters such as:
- Company details (name, location, industry)
- Company metrics (employee size, followers, revenue)
- Funding information (amount range)
- Founded year range
- Products/services
- School/university filter
Search for local businesses using Google Maps data with filters including:
- Business name
- Location (country, state, city)
- Industry
- Includes detailed information like ratings, reviews, contact details, and social media
$ pnpm installCreate a .env.development or .env.production file with your CUFinder API key:
CUFINDER_API_KEY=your_api_key_here# development
$ pnpm run start
# watch mode
$ pnpm run start:dev
# production mode
$ pnpm run start:prod# unit tests
$ pnpm run test
# e2e tests
$ pnpm run test:e2e
# test coverage
$ pnpm run test:covThe server exposes the following MCP-compatible endpoints:
POST /mcp/message- Main MCP message handling endpointPOST /mcp/sse- Server-Sent Events for streaming support
src/
├── tools/ # MCP tool implementations
│ ├── peoples.service.ts # PSE API integration
│ ├── companies.service.ts # CSE API integration
│ ├── lbs.service.ts # LBS API integration
│ └── tools.service.ts # MCP server setup
├── http/ # HTTP client with CUFinder API integration
├── shared/ # Shared types and models
│ ├── models/ # Data models
│ └── types/ # TypeScript interfaces
└── main.ts # Application entry point
MIT