mcp-server-claude

AmaanSarfaraz/mcp-server-claude

3.2

If you are the rightful owner of mcp-server-claude and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

The ConnectSafely API - MCP Server is a Model Context Protocol server designed to facilitate LinkedIn automation through a standardized interface.

Tools
5
Resources
0
Prompts
0

ConnectSafely API - MCP Server

An MCP (Model Context Protocol) server that provides tools for interacting with the ConnectSafely LinkedIn API. This server enables LinkedIn automation capabilities through a standardized MCP interface.

Table of Contents

Features

  • LinkedIn Automation: Send messages, connection requests, follow users, and more
  • Post Management: React, comment, and search LinkedIn posts
  • Profile Management: Fetch profile information and check relationships
  • Group Management: Get group members and interact within LinkedIn groups
  • Dual Transport Support: Works with both stdio and HTTP transports
  • Type-Safe: Built with TypeScript and runtime validation via Zod

Prerequisites

  • Node.js: Version 20.0.0 or higher
  • npm: Version 7 or higher (comes with Node.js)
  • ConnectSafely API Key: Get your API key from ConnectSafely

Installation

  1. Clone or navigate to the project directory:

    cd mcp-server-claude

  2. Install dependencies:

    npm install

  3. Build the project:

    npm run build

Configuration

Environment Variables

Create a .env file in the project root directory with your API credentials:

# Bearer Token Authentication (Required for most tools)
BEARER_TOKEN_BEARERAUTH=your_api_key_here

Getting Your API Key

  1. Sign up or log in to ConnectSafely
  2. Navigate to your API settings
  3. Generate or copy your API key
  4. Add it to your .env file as shown above

Optional: OAuth2 Configuration

If you need OAuth2 authentication (for certain advanced features), add these to your .env:

OAUTH_CLIENT_ID_SCHEMENAME=your_client_id
OAUTH_CLIENT_SECRET_SCHEMENAME=your_client_secret
OAUTH_SCOPES_SCHEMENAME=required_scopes

Running the Server

Stdio Transport (Default)

For use with MCP clients that communicate via stdio:

npm start

This will:

  1. Build the project automatically (via prestart script)
  2. Start the MCP server on stdio

HTTP Transport

For use with HTTP-based MCP clients:

npm run start:http

The server will start on http://localhost:3000 with:

  • MCP Endpoint: POST http://localhost:3000/mcp
  • Health Check: GET http://localhost:3000/health

Development Mode

For development with automatic rebuilding:

# Type check only
npm run typecheck

# Build only
npm run build

Available Tools

The server provides the following LinkedIn automation tools:

Connection & Messaging

  • follow-user: Follow or unfollow a LinkedIn user
  • send-message: Send a message to a LinkedIn user (supports group messaging)
  • send-connection-request: Send a connection request with optional custom message
  • check-relationship: Check relationship status with a LinkedIn user
  • check-relationship-specific-account: Check relationship using a specific account

Posts & Engagement

  • get-latest-posts: Get the latest posts from a LinkedIn user
  • react-to-post: React to a LinkedIn post (LIKE, PRAISE, APPRECIATION, etc.)
  • comment-on-post: Comment on a LinkedIn post
  • get-post-comments: Get comments from a post with pagination
  • get-all-post-comments: Get all comments with automatic pagination
  • search-posts: Search LinkedIn posts by keywords with filters
  • scrape-post: Scrape public LinkedIn post content (no auth required)

Profile & Organization

  • fetch-profile: Fetch detailed LinkedIn profile information
  • get-linkedin-organizations: Retrieve LinkedIn organizations/company pages you manage

Groups

  • get-group-members: Fetch members of a LinkedIn group by group ID
  • get-group-members-by-url: Fetch group members using the group URL

Comments

  • reply-to-comment: Reply to a comment on a LinkedIn post (not yet implemented)

Usage Examples

Example 1: Send a Connection Request

{
  "tool": "send-connection-request",
  "arguments": {
    "requestBody": {
      "profileId": "john-doe",
      "customMessage": "Hi John, I'd like to connect with you!"
    }
  }
}

Example 2: Search for Posts

{
  "tool": "search-posts",
  "arguments": {
    "requestBody": {
      "keywords": "TypeScript",
      "count": 10,
      "datePosted": "past-week",
      "sortBy": "relevance",
      "authorJobTitles": ["CTO", "Senior Developer"]
    }
  }
}

Example 3: Comment on a Post

{
  "tool": "comment-on-post",
  "arguments": {
    "requestBody": {
      "postUrl": "https://www.linkedin.com/posts/...",
      "comment": "Great insights! Thanks for sharing.",
      "tagPostAuthor": true
    }
  }
}

Example 4: Get Group Members

{
  "tool": "get-group-members",
  "arguments": {
    "requestBody": {
      "groupId": "123456",
      "count": 50,
      "membershipStatuses": ["MEMBER", "MANAGER"]
    }
  }
}

Example 5: Fetch Profile Information

{
  "tool": "fetch-profile",
  "arguments": {
    "requestBody": {
      "profileId": "jane-smith",
      "includeContact": true,
      "includeGeoLocation": false
    }
  }
}

Troubleshooting

Server Won't Start

  1. Check Node.js version:

    node --version  # Should be >= 20.0.0

  2. Verify dependencies are installed:

    npm install

  3. Check for build errors:

    npm run build

Authentication Errors

  1. Verify your API key is set:

    • Check that .env file exists in the project root
    • Ensure BEARER_TOKEN_BEARERAUTH is set correctly
    • No quotes needed around the API key value

  2. Check API key validity:

    • Verify the key is active in your ConnectSafely dashboard
    • Ensure there are no extra spaces or newlines

HTTP Transport Issues

  1. Port already in use:

    • The default port is 3000
    • Change the port in src/index.ts if needed (line 686)

  2. CORS errors:

    • CORS is enabled by default for all routes
    • Check browser console for specific CORS error messages

Tool Execution Errors

  1. Validation errors:

    • Check the tool's input schema requirements
    • Ensure all required fields are provided
    • Verify data types match (strings, numbers, booleans)

  2. API errors:

    • Check the error message for specific API response details
    • Verify your account has the necessary permissions
    • Some tools may require specific LinkedIn account types

Development

Project Structure

mcp-server-claude/
├── src/
│   ├── index.ts              # Main MCP server and tool definitions
│   └── streamable-http.ts    # HTTP transport implementation
├── build/                    # Compiled JavaScript (generated)
├── public/                   # Static files for web interface
├── package.json              # Dependencies and scripts
├── tsconfig.json             # TypeScript configuration
└── .env                      # Environment variables (create this)

Adding New Tools

  1. Add tool definition to toolDefinitionMap in src/index.ts
  2. Include proper JSON Schema for input validation
  3. Specify security requirements
  4. Rebuild: npm run build

Testing

Run type checking:

npm run typecheck

Code Style

  • TypeScript strict mode enabled
  • ES modules (type: "module")
  • Target: ES2022, Node16 module resolution
  • All errors logged to stderr (console.error)

Security Notes

  • Never commit your .env file to version control
  • API keys are sensitive credentials - keep them secure
  • The server validates all inputs using Zod schemas
  • Security schemes are automatically applied based on tool requirements

Support

For issues related to:

License

This project is private and proprietary.

Version: 1.0.0
Generated: 2025-11-21
API Base URL: https://api.connectsafely.ai