REPOSITORY OVERVIEW

Live repository statistics

★ 44Stars

⑂ 6Forks

◯ 15Open issues

◉ 44Watchers

70/100

OPENREPOHUB HEALTH SIGNAL

Healthy signals

A transparent discovery signal based on current public GitHub metadata.

Recent activity35% weight

100

Community adoption25% weight

Maintenance state20% weight

License clarity10% weight

100

Project information10% weight

100

This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.

README preview

LinkedIn MCP Server

📚 View Full Documentation & Installation Guides →

A comprehensive Model Context Protocol (MCP) server for LinkedIn API integration. Manage your LinkedIn profile, posts, connections, skills, education, certifications, and more through AI agents like Claude, ChatGPT, and other LLM applications.

🚀 Features

📱 Social Features

Profile Management: Fetch and view your LinkedIn profile
Posts & Engagement: Retrieve posts with engagement metrics, share new content
Connections: Get and manage your professional network
People Search: Find professionals by keywords

📝 Profile Management (18 Tools Total)

Skills: Add and remove skills from your profile
Work Experience: Add, update, and delete positions
Education: Manage educational background
Certifications: Add and remove professional certifications
Publications: Manage your published works
Languages: Add language proficiency to your profile

💻 Developer Experience

Modern MCP SDK: Built with latest McpServer API (v1.1.0+)
OpenID Connect Support: Works with LinkedIn's modern OAuth 2.0 + OIDC authentication
Full TypeScript support with strict type checking
Comprehensive test suite: 67 test cases, 85%+ server coverage
Zod schema validation for type safety and input validation
Modern async/await patterns
Extensive logging and error handling
Latest dependencies (Vitest 4, Zod 4, MCP SDK 1.24+)

📋 Prerequisites

Node.js >= 18.0.0
LinkedIn Developer App (Client ID & Secret) or existing access token
pnpm, npm, or yarn

📦 Installation

# Using pnpm (recommended)
pnpm install @pegasusheavy/linkedin-mcp

# Using npm
npm install @pegasusheavy/linkedin-mcp

# Using yarn
yarn add @pegasusheavy/linkedin-mcp

🔧 Configuration

Authentication Setup

This server requires LinkedIn API access. Choose your authentication method:

Option 1: Automatic OAuth Flow (Recommended) 🚀

The server automatically handles OAuth authentication when you don't have an access token.

Step 1: Create a LinkedIn App

Go to LinkedIn Developers
Click "Create App" and fill in the required information
Note your Client ID and Client Secret

Step 2: Configure OAuth Settings

In your app settings, go to "Auth" tab
Add http://localhost:50001/callback to "Authorized redirect URLs for your app"
Request the following Products (in Products tab):
- Sign In with LinkedIn using OpenID Connect (required for profile access)
- Share on LinkedIn (required for posting)

Note: The server uses OpenID Connect scopes (openid, profile, email, w_member_social) which work with the standard "Sign In with LinkedIn" product. No special API access required!

Step 3: Configure Your MCP Client

The server receives configuration through environment variables from your MCP client:

For Claude Desktop:

Edit your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Linux: ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id_here",
        "LINKEDIN_CLIENT_SECRET": "your_client_secret_here"
      }
    }
  }
}

For Cursor IDE:

In .cursor/mcp.json:

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id_here",
        "LINKEDIN_CLIENT_SECRET": "your_client_secret_here"
      }
    }
  }
}

What Happens:

First time the server starts, it detects no access token
Automatically opens your browser to LinkedIn's authorization page
You authorize the application once
Token is cached in memory for the session
Token is automatically refreshed when it expires (within the same session)
Each new session requires re-authorization (secure, no disk storage)

Option 2: Manual Access Token

If you already have an access token:

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "your_token_here"
      }
    }
  }
}

Configuration Options

Environment Variable	Required	Description
`LINKEDIN_CLIENT_ID`	For OAuth	Your LinkedIn app client ID
`LINKEDIN_CLIENT_SECRET`	For OAuth	Your LinkedIn app client secret
`LINKEDIN_REDIRECT_URI`	Optional	OAuth callback URL (default: `http://localhost:50001/callback`)
`LINKEDIN_ACCESS_TOKEN`	Alternative	Use existing token instead of OAuth
`LOG_LEVEL`	Optional	Logging verbosity: `debug`, `info`, `warn`, `error` (default: `info`)

Token Management

✨ Automatic Features:

In-Memory Caching - Tokens cached in memory during the session
Auto-Refresh - Expired tokens automatically refreshed within session
Session-Based - Authenticate once per MCP client session
No Disk Storage - Tokens never written to disk for maximum security

⚠️ Security Notes:

Memory only - Tokens stored in process memory, never on disk
Session scoped - Each new session requires re-authorization
CSRF protection - State parameter validation prevents attacks
Secure secrets - Client secrets never exposed to browser
Auto-cleanup - Tokens cleared when server stops

🎯 Usage

Claude Desktop

The easiest way to use this MCP server is with Claude Desktop.

Quick Start with OAuth (Recommended)

Create a LinkedIn App (see Configuration above)
Configure Claude Desktop

Open your Claude Desktop configuration file:
- macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
- Windows: %APPDATA%\Claude\claude_desktop_config.json
- Linux: ~/.config/Claude/claude_desktop_config.json
Add this configuration with your LinkedIn app credentials:
```
{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id_here",
        "LINKEDIN_CLIENT_SECRET": "your_client_secret_here"
      }
    }
  }
}
```
Restart Claude Desktop

Completely quit and restart Claude Desktop.
Authorize on First Use

The first time Claude tries to use LinkedIn tools:
- A browser window will open to LinkedIn's authorization page
- Click "Allow" to authorize the application
- The token is cached and automatically refreshed
- All future uses work seamlessly without re-authorization

Alternative: Use an Existing Token

If you already have a LinkedIn access token:

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "your_token_here"
      }
    }
  }
}

Step 4: Verify Installation

Once Claude Desktop restarts, you should see the LinkedIn MCP server connected. You can verify by asking Claude:

"Can you show me my LinkedIn profile?"
"What are my recent LinkedIn posts?"
"Add TypeScript to my LinkedIn skills"

Claude will now have access to all 18 LinkedIn tools! 🎉

Other MCP Clients

Cursor IDE

Add to your Cursor MCP settings (accessible via Cursor Settings → Features → Model Context Protocol):

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "your_linkedin_access_token_here"
      }
    }
  }
}

Cline (VS Code Extension)

Add to your Cline MCP settings:

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "your_linkedin_access_token_here"
      }
    }
  }
}

Continue

Add to ~/.continue/config.json:

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@pegasusheavy/linkedin-mcp"],
      "env": {
        "LINKEDIN_ACCESS_TOKEN": "your_linkedin_access_token_here"
      }
    }
  }
}

As a Standalone Server (Advanced)

import { LinkedInMCPServer } from '@pegasusheavy/linkedin-mcp';
import { getConfig, validateConfig } from '@pegasusheavy/linkedin-mcp/config';

const config = getConfig();
validateConfig(config);

const server = new LinkedInMCPServer(config);
await server.start();

With MCP Client (Advanced)

import { Client } from '@modelcontextprotocol/sdk/client/index.js';
import { StdioClientTransport } from '@modelcontextprotocol/sdk/client/stdio.js';

const transport = new StdioClientTransport({
  command: 'linkedin-mcp',
});

const client = new Client({
  name: 'my-app',
  version: '1.0.0',
}, {
  capabilities: {},
});

await client.connect(transport);

// List available tools
const tools = await client.listTools();

// Call a tool
const result = await client.callTool({
  name: 'get_linkedin_profile',
  arguments: {},
});

Using CLI

# Start the server
linkedin-mcp

# With environment variables
LINKEDIN_ACCESS_TOKEN=xxx linkedin-mcp

🛠️ Available Tools

Social & Content Tools

`get_linkedin_profile`

Get your LinkedIn profile information.

Arguments: None

Returns:

{
  "id": "user-id",
  "firstName": "John",
  "lastName": "Doe",
  "headline": "Software Engineer at Tech Corp",
  "vanityName": "johndoe"
}

`get_linkedin_posts`

Get your recent LinkedIn posts with engagement metrics.

Arguments:

limit (number, optional): Maximum number of posts (default: 10)

`get_linkedin_connections`

Get your LinkedIn connections.

Arguments:

limit (number, optional): Maximum connections (default: 50)

`share_linkedin_post`

Share a new post on LinkedIn.

Arguments:

text (string, required): The post content

Returns:

{
  "id": "post-id",
  "url": "https://www.linkedin.com/feed/update/post-id"
}

`search_linkedin_people`

Search for people on LinkedIn.

Arguments:

keywords (string, required): Search keywords
limit (number, optional): Maximum results (default: 10)

Profile Management Tools

Skills

add_linkedin_skill - Add a skill to your profile

name (string, required): Skill name

delete_linkedin_skill - Remove a skill

skillId (string, required): Skill ID to delete

Work Experience

add_linkedin_position - Add a position

title (string, required): Job title
company (string, required): Company name
startYear (number, required): Start year
startMonth (number, optional): Start month (1-12)
endYear (number, optional): End year
endMonth (number, optional): End month
description (string, optional): Job description
current (boolean, optional): Is current position?

**`u

ALGORITHMICALLY RELATED

Similar Open-Source Projects

Selected from shared topics, language and repository description—not editorial ratings.

Init-Abdellm /

AAABUILDER

★ 3

The first of a kind "AGENT AS A BACKEND" : Advanced AI/ML Agent Framework with TypeScript support, multi-model capabilities, vision controls, and comprehensive AI/ML model support including LLM, SLM, MLM, Vision, ASR, TTS, RL, GNN, RNN, CNN, GAN, Diffusion, Transformer, MLP, Autoencoder, BERT, RAG, Hybrid, Foundation, and fine-tuning

60/100 health

Active repositoryHas homepage

TypeScriptApache-2.0#aaab#agent-as-a-backend#agent-framework#agentframework

⑂ 0 forks◯ 0 issuesUpdated Jan 25, 2026

Project homepage ↗

View details View on GitHub ↗