mcp-servers/LINKEDIN_QUICK_START.md

2.3 KiB

LinkedIn MCP Quick Start

What Changed

The LinkedIn MCP server is Python-based, not Node.js. The Dockerfile has been updated accordingly.

Quick Steps

1. Create LinkedIn Profile (Do This FIRST)

On your Docker host machine (TrueNAS):

# Install the package
pip install linkedin-scraper-mcp

# Create profile by logging in
linkedin-mcp-server --login
# Or: linkedin-scraper-mcp --login

This creates ~/.linkedin-mcp/profile/ with your authenticated session.

2. Build the Docker Image

From your gateway directory:

cd /mnt/NVME/Docker/mcp-gateway
docker compose build linkedin-mcp

3. Start the Service

docker compose up linkedin-mcp

4. Verify It's Running

docker ps | grep mcp-linkedin
docker logs mcp-linkedin

Important Files

  • Dockerfile — Python 3.12 with LinkedIn MCP installed from GitHub
  • SETUP.md — Detailed setup guide
  • .env — Contains LINKEDIN_LOG_LEVEL and other config

Environment Variables

Already configured in your .env:

LINKEDIN_LOG_LEVEL=INFO
LINKEDIN_TIMEOUT=10000
LINKEDIN_CHROME_PATH=/usr/bin/chromium-browser

Docker Compose Entry

Already added to your docker-compose.yml:

linkedin-mcp:
  build:
    context: ./linkedin-mcp
    dockerfile: Dockerfile
  container_name: mcp-linkedin
  restart: unless-stopped
  environment:
    - LOG_LEVEL=INFO
    - CHROME_PATH=/usr/bin/chromium-browser
    - TIMEOUT=10000
  volumes:
    - ~/.linkedin-mcp/profile:/root/.linkedin-mcp/profile
  networks: [mcpnet]

Troubleshooting

Build Fails

Ensure you have:

  • Git installed
  • Internet access
  • Python 3.12+ on Docker host

Profile Issues

# Remove old profile
rm -rf ~/.linkedin-mcp/profile

# Create new one
linkedin-mcp-server --login

# Restart Docker service
docker compose restart linkedin-mcp

Check Logs

docker logs mcp-linkedin -f

Full Stack Start

Now you can start everything with one command:

docker compose up

This will start:

  • Gateway proxy (port 4444)
  • ERPNext MCP
  • TrueNAS MCP
  • Home Assistant MCP
  • Wave Finance MCP
  • LinkedIn MCP

Next: Dashboard

Once all services are running, access the dashboard at:

http://localhost:4444/dashboard

Or on your remote instance:

https://mcp.wilddragon.net/dashboard