LinkedIn Intelligence MCP Server

LinkedIn Intelligence MCP Server

Transform LinkedIn from a social network into your AI-powered business intelligence platform.

The LinkedIn MCP Server connects Claude Desktop directly to LinkedIn's data layer, enabling natural language conversations that extract insights, automate engagement, and supercharge your professional networking—all without leaving your AI assistant.

Why LinkedIn MCP?

The MCP Ecosystem Advantage

LinkedIn MCP becomes exponentially more powerful when combined with other Claude Desktop integrations:

• Gmail MCP → "Find LinkedIn profiles for everyone who emailed me this week about partnerships"
• Google Calendar MCP → "Who am I meeting tomorrow? Pull their LinkedIn profiles and recent posts"
• Notion/Obsidian MCP → "Add this prospect's LinkedIn summary to my CRM notes"
• File System MCP → "Export my network analytics to a CSV for my quarterly report"

This isn't just another LinkedIn tool—it's the bridge that connects your professional network to your entire AI-powered workflow.

Use Cases by Role

🎯 Sales Development Representatives (SDRs)

Turn LinkedIn into your AI-powered prospecting engine.

Your Daily Workflow, Supercharged

Key Tools for SDRs

🔍 Research & Intelligence
├── search_people() - Find prospects by company, title, keywords
├── get_profile() - Deep-dive into prospect backgrounds
├── get_profile_posts() - See what they care about
└── batch_get_profiles() - Research entire account lists at once

📬 Outreach & Engagement
├── send_connection_request() - Personalized invites at scale
├── send_message() - Direct outreach to connections
├── like_post() / comment_on_post() - Warm up cold prospects
└── react_to_post() - Show genuine engagement

📊 Pipeline Intelligence
├── get_conversations() - Track outreach threads
├── analyze_engagement() - See what messaging resonates
└── get_pending_invitations() - Manage your connection pipeline

Example Prompts for SDRs

• "Search for VPs of Engineering at Series B fintech companies in NYC"
• "What has [prospect name] posted about in the last month? Summarize the themes"
• "Send connection requests to these 10 people with a message about [topic]"
• "Show me which of my connection requests from last week were accepted"
• "Cross-reference my Gmail contacts with LinkedIn—who should I connect with?"

🤝 Business Development & Partnerships

Map relationships and identify partnership opportunities at scale.

Strategic Intelligence Gathering

Key Tools for BD Professionals

🏢 Company Intelligence
├── search_companies() - Find potential partners by criteria
├── get_profile_posts() - Track company announcements
└── search_people() - Identify key stakeholders

🔗 Relationship Mapping
├── get_connections() - Your network inventory
├── get_network_stats() - Network composition analysis
└── get_profile_contact_info() - Direct contact details

📈 Opportunity Tracking
├── get_feed() - Monitor partnership signals
├── analyze_content_performance() - Understand what resonates
└── generate_engagement_report() - Track relationship building

Example Prompts for BD

• "Find AI/ML companies that have posted about seeking partnerships in the last 30 days"
• "Map my path to the CEO of [target company]—who do I know there?"
• "What integrations or partnerships has [company] announced this quarter?"
• "Get me the profiles of everyone with 'Partnerships' in their title at these 5 companies"
• "Summarize the recent posts from decision-makers at my target accounts"

📈 Personal Brand Builders

Data-driven content strategy for thought leadership.

Build Your Presence Strategically

Key Tools for Brand Builders

📝 Content Creation
├── create_post() - Publish directly from Claude
├── create_draft() - Save ideas for later
├── analyze_draft_content() - Get AI feedback before posting
└── schedule_post() - Queue content for peak times

📊 Performance Analytics
├── analyze_optimal_posting_times() - Find your best windows
├── analyze_hashtag_performance() - Optimize discoverability
├── analyze_content_performance() - Track what works
└── generate_engagement_report() - Full performance review

🎯 Audience Intelligence
├── get_post_comments() - Understand audience reactions
├── get_post_reactions() - See who engages
├── analyze_post_audience() - Demographics of engagers
└── analyze_engagement() - Deep engagement metrics

Example Prompts for Brand Builders

• "Analyze my last 30 posts—which topics drove the most engagement?"
• "What time of day should I post to maximize reach with my audience?"
• "Draft a post about [topic] optimized for engagement based on what works for me"
• "Schedule this post for my optimal posting time tomorrow"
• "Which of my posts this month could be repurposed into a series?"

🏢 Agency & Competitive Intelligence

Multi-account insights and competitor analysis for agency professionals.

Agency-Grade Intelligence

Key Tools for Agencies

🔍 Competitive Analysis
├── get_profile_posts() - Track competitor content
├── analyze_content_performance() - Benchmark strategies
├── analyze_engagement() - Compare engagement metrics
└── search_people() - Monitor competitor teams

📋 Client Management
├── generate_engagement_report() - Automated reporting
├── analyze_optimal_posting_times() - Client-specific insights
├── analyze_hashtag_performance() - Strategy optimization
└── get_post_analytics() - Detailed post breakdowns

📈 Market Intelligence
├── get_feed() - Industry trend monitoring
├── search_companies() - Market landscape mapping
└── batch_get_profiles() - Bulk research capabilities

Example Prompts for Agencies

• "Compare the posting strategy of [client] vs their top 3 competitors"
• "Generate a monthly engagement report for [client profile]"
• "What content themes are driving the most engagement in the [industry] space?"
• "Track what [competitor] has posted about our client's product category"
• "Create a benchmark report: my client vs industry average engagement"

👥 Recruiters & Talent Acquisition

Build talent pipelines with AI-powered sourcing.

Recruiting, Reimagined

Key Tools for Recruiters

🎯 Candidate Sourcing
├── search_people() - Advanced candidate search
├── get_profile() - Comprehensive candidate profiles
├── get_profile_skills() - Skills and endorsements
├── batch_get_profiles() - Research candidate lists

💬 Candidate Engagement
├── get_profile_posts() - Understand candidate interests
├── send_connection_request() - Personalized outreach
├── send_message() - Direct candidate communication
└── get_conversations() - Track outreach threads

📊 Pipeline Analytics
├── get_pending_invitations() - Monitor connection pipeline
├── get_connections() - Your talent network
└── get_network_stats() - Network composition

Example Prompts for Recruiters

• "Find product managers at [company] who've been in role 2+ years"
• "What has this candidate posted about? What motivates them?"
• "Get profiles for all engineers at [company] with ML in their title"
• "Draft a personalized connection message for [candidate] based on their interests"
• "Who in my network is connected to this candidate for a warm intro?"

✍️ Content Creators & Thought Leaders

Maximize reach and engagement with data-driven content strategies.

Content Strategy on Autopilot

Key Tools for Creators

📝 Content Workflow
├── create_draft() - Save ideas and drafts
├── analyze_draft_content() - AI-powered content review
├── schedule_post() - Automated publishing
└── list_scheduled_posts() - Manage content calendar

📊 Deep Analytics
├── analyze_content_performance() - Content type analysis
├── analyze_hashtag_performance() - Hashtag ROI
├── analyze_optimal_posting_times() - Peak engagement windows
└── generate_engagement_report() - Comprehensive analytics

🎯 Engagement Optimization
├── get_post_comments() - Comment analysis
├── reply_to_comment() - Community management
├── analyze_post_audience() - Audience demographics
└── get_post_reactions() - Reaction breakdown

Example Prompts for Creators

• "Analyze my content performance over the last 90 days—what's working?"
• "Which of my posts should I turn into a content series?"
• "What are the top creators in [niche] posting about this month?"
• "Review this draft and suggest improvements for engagement"
• "Build me a content calendar for next week based on my best-performing topics"

Quick Start

Prerequisites

• Python 3.11+
• uv (recommended) or pip
• Claude Desktop
• LinkedIn account

Installation

# Clone the repository
git clone https://github.com/southleft/linkedin-mcp.git
cd linkedin-mcp

# Set up environment
uv venv && source .venv/bin/activate
uv pip install -e ".[dev]"

# Create data directory
mkdir -p data

Authentication

Export your LinkedIn cookies from an active browser session:

1. Log into LinkedIn in Chrome/Firefox
2. Open DevTools → Application (Chrome) or Storage (Firefox)
3. Find Cookies → https://www.linkedin.com
4. Copy li_at and JSESSIONID values

Create data/session_cookies.json:

{
  "li_at": "YOUR_LI_AT_COOKIE_VALUE",
  "JSESSIONID": "\"ajax:1234567890123456789\""
}

Note: The JSESSIONID typically includes escaped quotes. Copy exactly as shown in your browser.

Claude Desktop Configuration

Add to your Claude Desktop config file:

{
  "mcpServers": {
    "linkedin": {
      "command": "uv",
      "args": ["run", "linkedin-mcp"],
      "cwd": "/absolute/path/to/linkedin-mcp",
      "env": {
        "LINKEDIN_API_ENABLED": "true",
        "LOG_LEVEL": "INFO"
      }
    }
  }
}

Important: The cwd must be an absolute path to the project directory.

Restart Claude Desktop and verify with: "Use debug_context to check LinkedIn MCP status"

Complete Tool Reference

View All 60+ Available Tools

Diagnostic

• debug_context() - Server initialization status and configuration

Profile Tools

• get_my_profile() - Your LinkedIn profile
• get_profile(profile_id) - Any profile by public ID
• get_profile_contact_info(profile_id) - Contact information
• get_profile_skills(profile_id) - Skills and endorsements
• get_profile_sections() - Editable profile sections
• get_profile_completeness() - Profile completeness score
• get_network_stats() - Network size and demographics
• batch_get_profiles(profile_ids) - Multiple profiles efficiently

Profile Management (browser automation)

• update_profile_headline(headline) - Update headline
• update_profile_summary(summary) - Update about section
• upload_profile_photo(photo_path) - Upload profile photo
• upload_background_photo(photo_path) - Upload banner photo
• add_profile_skill(skill_name) - Add a skill
• check_browser_automation_status() - Browser availability

Feed & Posts

• get_feed(limit) - Your feed posts
• get_profile_posts(profile_id, limit) - Posts from a profile
• create_post(text, visibility) - Create a new post
• get_post_reactions(post_urn) - Reactions on a post
• get_post_comments(post_urn) - Comments on a post

Content Creation & Scheduling

• analyze_draft_content(content, industry) - Content analysis with suggestions
• create_draft(content, title, tags) - Save a content draft
• list_drafts(tag) - List content drafts
• get_draft(draft_id) - Get a specific draft
• update_draft(draft_id, content, title, tags) - Update a draft
• delete_draft(draft_id) - Delete a draft
• publish_draft(draft_id, visibility) - Publish a draft as a post
• schedule_post(content, scheduled_time, visibility, timezone) - Schedule a post
• list_scheduled_posts(status) - View scheduled posts
• get_scheduled_post(job_id) - Scheduled post details
• update_scheduled_post(job_id, ...) - Update scheduled post
• cancel_scheduled_post(job_id) - Cancel scheduled post

Engagement

• like_post(post_urn) - Like a post
• react_to_post(post_urn, reaction) - React (LIKE, CELEBRATE, SUPPORT, LOVE, INSIGHTFUL, FUNNY)
• unreact_to_post(post_urn) - Remove reaction
• comment_on_post(post_urn, text) - Comment on a post
• reply_to_comment(comment_urn, text) - Reply to a comment

Messaging

• get_conversations(limit) - Messaging conversations
• get_conversation(conversation_id) - Full conversation history
• send_message(profile_id, text) - Send a direct message
• send_bulk_messages(profile_ids, text) - Send to multiple recipients (max 25)

Connections

• get_connections(limit) - Your connections
• send_connection_request(profile_id, message) - Send connection invite
• remove_connection(profile_id) - Remove a connection
• get_pending_invitations(sent) - View pending invites
• accept_invitation(invitation_id, shared_secret) - Accept connection request
• reject_invitation(invitation_id, shared_secret) - Reject connection request

Search

• search_people(keywords, limit, connection_of, current_company) - Search for people
• search_companies(keywords, limit) - Search for companies

Analytics

• get_post_analytics(post_urn) - Engagement metrics for a post
• analyze_engagement(post_urn, follower_count) - Deep engagement analysis
• analyze_content_performance(profile_id, post_limit) - Content type analysis
• analyze_optimal_posting_times(profile_id, post_limit) - Best times to post
• analyze_post_audience(post_urn) - Audience demographics from commenters
• analyze_hashtag_performance(profile_id, post_limit) - Hashtag effectiveness
• generate_engagement_report(profile_id, post_limit) - Comprehensive engagement report
• get_rate_limit_status() - API rate limit status
• get_cache_stats() - Cache performance statistics

Environment Variables

Troubleshooting

1. Ensure LINKEDIN_API_ENABLED is set to "true" (as a string) in Claude Desktop config
2. Verify data/session_cookies.json exists with valid cookies
3. Check that cookies haven't expired (re-export from browser if needed)
4. Verify the cwd path is absolute, not relative

LinkedIn session cookies typically last 1-2 weeks. If you're getting authentication errors:

1. Log into LinkedIn in your browser
2. Re-export the li_at and JSESSIONID cookies
3. Update data/session_cookies.json
4. Restart Claude Desktop

# macOS
tail -f ~/Library/Logs/Claude/mcp-server-linkedin.log

# Windows (PowerShell)
Get-Content -Path "$env:APPDATA\Claude\logs\mcp-server-linkedin.log" -Wait

# Linux
tail -f ~/.config/Claude/logs/mcp-server-linkedin.log

Architecture

linkedin-mcp/
├── src/linkedin_mcp/
│   ├── config/          # Settings and constants
│   ├── core/            # Context, exceptions, logging, lifespan
│   ├── db/              # SQLAlchemy models and repositories
│   ├── services/
│   │   ├── linkedin/    # LinkedIn API wrapper
│   │   ├── browser/     # Playwright automation
│   │   ├── scheduler/   # APScheduler integration
│   │   └── analytics/   # Analytics processing
│   ├── tools/           # MCP tool definitions
│   ├── resources/       # MCP resource definitions
│   ├── prompts/         # MCP prompt templates
│   ├── server.py        # FastMCP server
│   └── main.py          # Entry point
├── scripts/             # Utility scripts
├── tests/               # Test suite
└── data/                # Runtime data (gitignored)

Security & Responsible Use

Important: This server uses the unofficial LinkedIn API which accesses LinkedIn's internal Voyager API.

• Use a secondary LinkedIn account for testing
• Respect rate limits (~900 requests/hour)
• Don't spam or automate engagement excessively
• Your account may be restricted if LinkedIn detects automation
• Never commit your data/ directory or cookie files

Development

# Run tests
pytest tests/ -v --cov=linkedin_mcp

# Linting
ruff check src/ && ruff format src/

# Type checking
mypy src/linkedin_mcp

# Run locally
uv run linkedin-mcp

License

MIT License - see LICENSE for details.

Acknowledgments

• tomquirk/linkedin-api - Unofficial LinkedIn API
• FastMCP - MCP SDK for Python
• Model Context Protocol - AI tool protocol

Built for professionals who move fast. Report Issues · Contribute