linkedin-post-mcp

A Model Context Protocol (MCP) server that lets AI agents (Claude, Cursor, etc.) create and manage LinkedIn posts on your behalf via the LinkedIn REST API.

Features

• 🔐 OAuth 2.0 Authentication — Secure 3-legged OAuth flow with automatic token refresh

• 📝 Text Posts — Create plain text posts on your personal profile or company page

• 🔗 URL/Article Shares — Share links with custom commentary

• 🗑️ Delete Posts — Remove posts programmatically

• 👤 Profile Lookup — Fetch your LinkedIn profile and URN

• 🏢 Company Page Support — Post as a company/organization page

Related MCP server: LinkedIn MCP Server

Prerequisites

1. LinkedIn Developer App — Create one here

   • Enable the "Share on LinkedIn" product

   • Enable "Sign In with LinkedIn using OpenID Connect"

   • Add http://localhost:8080/callback as an authorized redirect URL

   • Copy your CLIENT_ID and CLIENT_SECRET

2. npm account — Sign up at npmjs.com (only needed to publish)

Installation

npx linkedin-post-mcp@latest

Or install globally:

npm install -g linkedin-post-mcp

Configuration

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "linkedin-post-mcp@latest"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id_here",
        "LINKEDIN_CLIENT_SECRET": "your_client_secret_here",
        "LINKEDIN_REDIRECT_URI": "http://localhost:8080/callback"
      }
    }
  }
}

Cursor / Other MCP Hosts

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

Authentication Flow

On first use:

1. Ask the AI: "Get the LinkedIn auth URL"

2. Open the URL in your browser and authorize the app

3. LinkedIn redirects you to http://localhost:8080/callback?code=XXX

4. Copy the code value and tell the AI: "Exchange this code: XXX"

5. Done! Tokens are stored in ~/.linkedin-mcp-tokens.json

Available Tools

Example Prompts

"Post on LinkedIn: Excited to announce our new product launch! #tech #startup"

"Share this article on LinkedIn with the comment 'Great read!': https://example.com/article"

"Create a LinkedIn post on behalf of our company page (org ID: 12345678): ..."

"Check my LinkedIn auth status"

"Delete my last LinkedIn post with URN urn:li:share:..."

Environment Variables

Rate Limits

LinkedIn's API allows approximately 100 API calls per day per member on standard access. The server does not implement rate limit tracking, so be mindful of how frequently you call the tools.

License

MIT