LinkedIn MCP Server

Programmable LinkedIn Automation via Playwright and the Model Context Protocol.

📌 Overview

This MCP server exposes LinkedIn as a programmable platform via Playwright browser automation. It provides 20 strictly typed tools across profile management, connections, messaging, feed interaction, and job searching. All interactions are driven by automating a real Chromium browser against LinkedIn's web UI, entirely bypassing unofficial APIs.

Related MCP server: LinkedIn MCP Server

🏗️ Architecture

graph TD
    Host[MCP Host / Claude Desktop] <-->|JSON-RPC over stdio| MCP[LinkedIn MCP Node.js Server]
    
    subgraph Core Automation Engine
        MCP --> SessionMgr[Session & Cookie Manager]
        MCP --> Tools[20 Typed Action Tools]
        Tools --> Playwright[Playwright API]
    end
    
    Playwright -->|CDP Commands| Chrome[Non-Headless Chromium]
    Chrome <-->|Web Traffic| LI[(LinkedIn Web Platform)]

🚀 Getting Started

Prerequisites

• Node.js 20+

• A LinkedIn Account

• Environment configuration

Setup

# Install dependencies
npm install

# Build the project
npm run build

# Configure environment
cp .env.example .env

Edit your .env to include dummy/example credentials or your real credentials if deploying locally:

LINKEDIN_EMAIL=example@example.com
LINKEDIN_PASSWORD=your_password
COOKIES_PATH=./linkedin_cookies.json

(Note: Session cookies are persisted automatically after the first successful login).

🛠️ Tool Ecosystem

👤 Profile Tools

• get_my_profile / get_my_profile_full

• get_person_profile / get_person_profile_full

✏️ Profile Editing (ProseMirror compatible)

• update_profile_text / edit_headline

• add_experience_block / add_skill

🤝 Networking Pipeline

• search_people / connect_with_person

• smart_connect: Autonomous multi-phase networking pipeline (Scrape -> Search -> Score -> Reach Out).

💬 Messaging & Feed

• get_inbox / send_message

• get_feed / get_company_posts / create_post

💼 Job Search

• search_jobs / get_job_details

⚙️ Technical Highlights

• Anti-Bot Circumvention: Utilizes non-headless Chromium instances to mimic legitimate user interaction.

• ProseMirror Handlers: Gracefully handles LinkedIn's complex TipTap/ProseMirror rich-text editors for post creation and profile updates.

• Dry-Run Safety: All mutating endpoints support a dryRun boolean, executing form logic while bypassing the final submit action.

• Session Resilience: Implements automated login detection and session recovery loops with explicit CAPTCHA abortion.