LinkedIn MCP Server

A Model Context Protocol (MCP) server for LinkedIn API integration, enabling LLMs to interact with LinkedIn's professional network.

Features

• Profile Management: Get your own profile, view other profiles
• Connection Management: List connections, send connection requests, check connection status
• Posts & Content: Create posts, get feed, like and comment on posts
• Job Search: Search and view job listings, apply insights
• Company Pages: View company info, followers, posts
• Messaging: Send messages, view conversations
• Search: Search for people, companies, jobs

Authentication

This server uses LinkedIn's OAuth 2.0. You need:

1. A LinkedIn Developer Application: https://www.linkedin.com/developers/
2. An OAuth 2.0 Access Token with the appropriate scopes

Required OAuth Scopes

Scope	Purpose
r_liteprofile	Read your basic profile
r_emailaddress	Read your email address
r_fullprofile	Read your full profile
w_member_social	Write posts, likes, comments
r_member_social	Read posts, likes, comments
r_1st_connections	Read your connections
w_messages	Send messages

Setup

1. Install

pip install linkedin-mcp
# or from source:
pip install -e .

2. Configure Environment

Create a .env file:

LINKEDIN_ACCESS_TOKEN=your_access_token_here
LINKEDIN_CLIENT_ID=your_client_id_here
LINKEDIN_CLIENT_SECRET=your_client_secret_here

3. Run the Server

# stdio mode (for Claude Desktop)
linkedin-mcp

# HTTP mode (for remote access)
linkedin-mcp --transport streamable_http --port 8000

4. Configure Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "linkedin": {
      "command": "linkedin-mcp",
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "your_access_token_here"
      }
    }
  }
}

Available Tools

Profile Tools

• linkedin_get_my_profile - Get your LinkedIn profile
• linkedin_get_profile - Get any LinkedIn profile by ID or URL
• linkedin_update_profile_headline - Update your headline

Connection Tools

• linkedin_get_connections - List your connections
• linkedin_send_connection_request - Send a connection request
• linkedin_get_connection_status - Check connection status

Post & Content Tools

• linkedin_create_post - Create a LinkedIn post
• linkedin_get_feed - Get your LinkedIn feed
• linkedin_like_post - Like a post
• linkedin_comment_on_post - Comment on a post
• linkedin_get_post_comments - Get comments on a post
• linkedin_delete_post - Delete your post

Job Tools

• linkedin_search_jobs - Search for jobs
• linkedin_get_job_details - Get details for a specific job
• linkedin_get_job_recommendations - Get personalized job recommendations

Company Tools

• linkedin_get_company - Get company information
• linkedin_get_company_posts - Get posts from a company
• linkedin_follow_company - Follow a company

Message Tools

• linkedin_send_message - Send a message
• linkedin_get_conversations - List your conversations
• linkedin_get_conversation_messages - Get messages in a conversation

Search Tools

• linkedin_search_people - Search for people
• linkedin_search_companies - Search for companies

Development

# Install dev dependencies
pip install -e ".[dev]"

# Test the server
python -m linkedin_mcp.server

# Run with MCP Inspector
npx @modelcontextprotocol/inspector python -m linkedin_mcp.server

Important Notes

• LinkedIn's API has strict rate limits. The server handles 429 responses gracefully.
• Some features require LinkedIn Partner Program access.
• Always comply with LinkedIn's Terms of Service and API usage policies.
• The w_member_social scope is required for posting and engagement actions.

License

MIT License - see LICENSE file for details.