LinkedIn Outreach MCP Server

By MEOK AI Labs | meok.ai

AI-powered content generation for LinkedIn outreach. Helps sales teams and recruiters craft personalized connection requests, InMails, posts, comments, and multi-touch outreach sequences.

This is a content assistant, not a scraper or bot. It does NOT interact with LinkedIn's API, scrape profiles, or automate any actions. It generates text that humans review, edit, and send manually. Fully compliant with LinkedIn's Terms of Service.

Tools

Installation

pip install mcp

No additional dependencies. Pure Python, no API keys required.

Usage

Run the server

python server.py

Claude Desktop config

{
  "mcpServers": {
    "linkedin-outreach": {
      "command": "python",
      "args": ["/path/to/linkedin-outreach-mcp/server.py"]
    }
  }
}

Example calls

Generate a connection request:

Tool: generate_connection_request
Input: {"name": "Sarah Chen", "title": "VP of Engineering", "company": "Stripe", "shared_interests": "AI infrastructure, developer tools"}
Output: 3 personalized variants, all under 300 characters

Generate an InMail (sales tone):

Tool: generate_inmessage
Input: {"recipient_name": "James Park", "recipient_title": "CTO", "goal": "book a demo of our observability platform", "tone": "sales"}
Output: Value-first sales message with soft CTA

Generate a LinkedIn post:

Tool: generate_post
Input: {"topic": "AI in sales enablement", "key_points": "personalization at scale, shorter sales cycles, better lead scoring", "style": "thought-leadership"}
Output: Engaging post with hook, numbered points, CTA, and hashtags

Analyze a profile before outreach:

Tool: analyze_profile
Input: {"profile_text": "VP of Engineering at Stripe. Building payments infrastructure for the internet economy. Previously Google, MIT CS. Passionate about developer experience and distributed systems."}
Output: Industry (Technology), Seniority (VP/Executive), pain points, conversation starters, interest areas

Generate a 5-touch outreach sequence:

Tool: generate_outreach_sequence
Input: {"target_name": "Maria Rodriguez", "target_title": "Head of Talent", "target_company": "Figma", "goal": "partner on recruiting automation"}
Output: 5 messages with timing (Day 1 -> Day 25), each with strategic context

Generate a thoughtful comment:

Tool: generate_comment
Input: {"post_content": "Unpopular opinion: cold outreach is dead. The future is warm introductions and community-led growth. Here's why..."}
Output: 2-3 insightful comments that add a new perspective, not generic agreement

How It Works

This server generates content using rule-based templates and industry heuristics. It does NOT call any external AI APIs or LinkedIn APIs. All content is generated locally and instantly.

The analyze_profile tool extracts signals from profile text using keyword matching to identify:

• Industry (Technology, Sales, Marketing, Finance, Healthcare, etc.)

• Seniority level (C-Suite, VP, Director, Senior, Manager, Junior)

• Likely pain points based on role and industry

• Tailored conversation starters

• Mutual interest areas for common ground

Compliance

• Does NOT access LinkedIn's API

• Does NOT scrape any LinkedIn data

• Does NOT automate sending messages

• Does NOT store any user data or profile information

• All content is generated locally -- nothing leaves your machine

• Humans review and send all messages manually

This tool is a writing assistant, similar to Grammarly or a copywriting template. It helps you write better LinkedIn messages faster.

Pricing

License

MIT