Captain Data MCP API

A Model Context Protocol (MCP) server for Captain Data tools, designed to work with ChatGPT and other AI assistants. This API provides a clean interface for LinkedIn data extraction and search capabilities through Captain Data's powerful scraping engine.

Features

• MCP-compliant API for seamless integration with AI assistants

• Dual authentication system - API key headers or session tokens

• Comprehensive tool suite for LinkedIn data extraction and search

• Production-ready with proper error handling, logging, and monitoring

• TypeScript-first with strict type checking and validation

• Fastify-powered for high performance and extensibility

• Redis integration for session management with in-memory fallback

• Interactive API documentation with OpenAPI/Swagger

• Comprehensive testing with Jest and integration tests

Related MCP server: LinkedIn MCP Server

Authentication

The API supports two authentication methods:

1. Direct API Key: Set your Captain Data API key in the X-API-Key header

2. Session Tokens: Use the /auth endpoint to get a session token, then include it in the Authorization: Bearer <token> header

Local Development

Prerequisites

• Node.js 16+

• npm or yarn

• Captain Data API key

• Redis (optional, falls back to in-memory storage)

Setup

1. Install dependencies:

2. Create environment configuration:

3. Start the development server:

The server will be available at http://localhost:3000.

Development Scripts

Production Deployment

Environment Variables

Production Checklist

Before deploying to production, ensure:

• Environment variables are properly configured

• All tests pass: npm test

• Type checking passes: npm run type-check

• Rate limiting is configured appropriately

• Logging level is set to info or warn

• Health check endpoint is monitored

• Error tracking is set up (Sentry recommended)

• Redis is configured for session management (optional)

Monitoring & Observability

The API provides comprehensive monitoring capabilities:

• Health Check: GET /health - Returns system status and uptime

• Request Logging: All requests are logged with unique request IDs

• Error Tracking: Structured error responses with detailed context

• Performance Metrics: Response times and execution metrics

• Sentry Integration: Automatic error reporting and performance monitoring

API Endpoints

Core Endpoints

• GET /health - Health check and system status

• GET /introspect - List available tools (basic mode)

• GET /introspect?v=full - List all available tools (full mode)

• POST /auth - Generate session token from API key

• POST /tools/:alias - Execute a specific tool

Documentation

• GET /docs - Interactive API documentation (Swagger UI)

• GET /docs/json - OpenAPI specification

• GET /openapi.json - Raw OpenAPI specification

• GET /openapi.gpt.json - GPT-compatible OpenAPI specification

Available Tools

Tool Naming Convention

People Tools

• find_person - Find a person's LinkedIn profile by name

  • Input: full_name (required), company_name (optional)

  • Returns: LinkedIn profile URL, UID, and profile ID

• search_people - Find prospects using Sales Navigator with advanced filtering

  • Input: query (Sales Navigator query param), page, page_size

  • Returns: List of matching profiles with pagination

• enrich_person - Extract detailed profile information from LinkedIn URLs

  • Input: li_profile_url (required), full_enrich (optional boolean)

  • Returns: Comprehensive profile data including experience, skills, and contact info

Company Tools

• find_company - Find a company's LinkedIn page by name

  • Input: company_name (required)

  • Returns: LinkedIn company URL, UID, and company ID

• search_companies - Discover target companies based on various criteria

  • Input: query (Sales Navigator query param), page, page_size

  • Returns: List of matching companies with pagination

• enrich_company - Get comprehensive company data from LinkedIn company pages

  • Input: li_company_url (required)

  • Returns: Company details, employee count, industry, locations, and more

• search_company_employees - Identify key contacts within specific organizations

  • Input: company_uid (required), page, page_size

  • Returns: List of employees with their profile information

Utility Tools

• get_quotas - Get current workspace quota and billing information

  • Input: None

  • Returns: Credits remaining, credits used, plan name, billing cycle dates

Pagination

Search endpoints return pagination headers for navigating through results:

Use the page parameter to manually paginate through results.

Using with AI Assistants

ChatGPT Integration

1. Deploy to Vercel:

2. Configure in ChatGPT:

   • Use the deployed URL: https://your-project.vercel.app

   • The API will be available for ChatGPT Actions

Custom GPT Authentication Workaround

Important: Custom GPTs have limitations with authorization headers. To work around this, the API supports authentication via query parameters:

Note: Query parameter authentication is provided as a workaround for Custom GPTs. For production applications, prefer header-based authentication for better security.

MCP Protocol Support

The API is fully compatible with the Model Context Protocol (MCP), providing:

• Standardized tool introspection

• Consistent error handling

• Session-based authentication

• Structured request/response formats

Project Architecture

Testing

The project includes comprehensive testing:

Test Structure

• Unit Tests: Individual function and module testing

• Integration Tests: Full API workflow testing

• Authentication Tests: Session token and API key validation

• Error Handling Tests: Comprehensive error scenario coverage

Error Handling

The API provides structured error responses with:

• Unique request IDs for tracking

• Standardized error codes for programmatic handling

• Detailed error messages with context

• HTTP status codes following REST conventions

• Sentry integration for error tracking in production

Performance & Scalability

• Fastify framework for high-performance request handling

• Redis integration for scalable session management

• Rate limiting to prevent abuse

• Request timeouts and retry logic

• Connection pooling for external API calls

• Graceful degradation with in-memory fallbacks

Security Features

• Input validation with JSON schemas

• Rate limiting to prevent abuse

• Security headers (CORS, XSS protection, etc.)

• Authentication validation for all protected endpoints

• Request logging with sensitive data redaction

• Error message sanitization to prevent information leakage

Contributing

1. Fork the repository

2. Create a feature branch

3. Make your changes

4. Add tests for new functionality

5. Ensure all tests pass

6. Submit a pull request

License

MIT License - see LICENSE file for details.