LinkedIn MCP Server

"Let AI Assistants Work With LinkedIn" by John Rippy | johnrippy.link

---

AI-Powered LinkedIn Automation Through MCP

Model Context Protocol (MCP) is an open standard that lets AI assistants like Claude use external tools. This server exposes LinkedIn functionality to your AI assistant.

What can you do with this?

• Scrape LinkedIn profiles and posts through natural conversation
• Train AI voice models on writing styles
• Generate content that matches someone's voice
• Analyze post performance and suggest topics
• Search for people and companies

Talk to Claude. Let Claude talk to LinkedIn.

---

What is MCP?

MCP (Model Context Protocol) is a standardized way for AI assistants to use external tools. Instead of copying and pasting data, you can just ask Claude:

"Analyze my LinkedIn profile and tell me what topics I should write about"

Claude uses this MCP server to scrape your profile, analyze your posts, and give you actionable recommendations.

---

Available Tools

Profile Tools

Tool	Description
scrape_profile	Get full LinkedIn profile (name, headline, about, experience, posts)
scrape_posts	Scrape posts from a profile or by hashtag/keyword search

Voice Training Tools

Tool	Description
analyze_voice	Analyze writing style, tone, hooks, CTAs, and vocabulary
generate_post	Generate new posts matching a trained voice profile
improve_draft	Improve your draft to better match a voice profile

Content Tools

Tool	Description
suggest_topics	Get topic suggestions based on voice profile and trends
analyze_performance	Analyze engagement metrics and patterns from posts

Search Tools

Tool	Description
search_people	Search LinkedIn profiles by keywords, company, location

---

Quick Start Examples

Example 1: Scrape a Profile

{
  "mode": "demo",
  "tool": "scrape_profile",
  "profileUrl": "https://www.linkedin.com/in/johnrippy/",
  "maxPosts": 20
}

Example 2: Analyze Voice Style

{
  "mode": "demo",
  "tool": "analyze_voice",
  "profileUrl": "https://www.linkedin.com/in/johnrippy/",
  "saveProfile": true,
  "profileName": "johnrippy-voice"
}

Example 3: Generate a Post in Someone's Voice

{
  "mode": "demo",
  "tool": "generate_post",
  "voiceProfileId": "johnrippy-voice",
  "topic": "Why I stopped checking email first thing in the morning",
  "contentType": "text",
  "includeHashtags": true
}

Example 4: Improve a Draft

{
  "mode": "demo",
  "tool": "improve_draft",
  "voiceProfileId": "johnrippy-voice",
  "draft": "Here are my thoughts on productivity...",
  "feedback": "Make it more engaging with a stronger hook"
}

---

Setup with Claude Desktop

Option 1: Using Apify (Cloud)

Add to your Claude Desktop config (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "linkedin": {
      "command": "npx",
      "args": ["-y", "@anthropic-ai/model-context-protocol", "run", "apify-actor", "localhowl/linkedin-mcp-server"],
      "env": {
        "APIFY_TOKEN": "your-apify-token"
      }
    }
  }
}

Option 2: Running Locally

{
  "mcpServers": {
    "linkedin": {
      "command": "node",
      "args": ["/path/to/linkedin-mcp-server/dist/main.js"],
      "env": {
        "LINKEDIN_COOKIES": "[...]",
        "ANTHROPIC_API_KEY": "sk-ant-..."
      }
    }
  }
}

---

Authentication

LinkedIn scraping requires authentication. Provide one of:

Method	Description
Cookies (recommended)	Export your LinkedIn session cookies as JSON
Email/Password	LinkedIn credentials for automated login

For AI features (voice training, post generation), provide:

• Claude API key (recommended)
• OpenAI API key (alternative)

---

Example Conversation with Claude

User: Can you analyze my LinkedIn voice and suggest some post topics?

Claude: I'll analyze your LinkedIn profile to understand your writing style.
[Uses scrape_posts tool]
[Uses analyze_voice tool]

Based on your posts, here's your voice profile:
- Tone: Professional but conversational
- Hooks: You often start with contrarian statements
- CTAs: You typically end with questions
- Signature phrases: "Here's the thing", "unpopular opinion"

Let me suggest some topics that match your style:
[Uses suggest_topics tool]

Here are 5 topic ideas:
1. "Why I stopped optimizing for likes" - matches your contrarian style
2. "The 5-minute habit that changed my morning" - fits your list format
3. "What nobody tells you about growing on LinkedIn"
4. "I deleted my content calendar. Here's what happened."
5. "The best career advice I ever ignored"

Want me to draft one of these?

---

Pay-Per-Event Pricing

You only pay for what you use.

Event	Description	Price
profile_scraped	Profile data extracted	$0.01
posts_scraped	Per batch of posts	$0.01-0.05
voice_analyzed	Voice profile created	$0.02
post_generated	Content generated	$0.01

Cost Examples

Task	Estimated Cost
Scrape profile with 20 posts	~$0.02
Analyze voice profile	~$0.02
Generate 5 post drafts	~$0.05
Full content strategy session	~$0.15

---

Use Cases

Personal Branding

• Voice consistency: Maintain your unique voice across all posts
• Content ideation: Get topic suggestions based on your style
• Draft improvement: Polish drafts to match your voice

Ghostwriting

• Client voice training: Learn a client's writing style
• Authentic content: Generate posts that sound like them
• Quality control: Ensure drafts match the trained voice

Content Agencies

• Scale content creation: Generate drafts for multiple clients
• Consistency: Maintain unique voices across accounts
• Research: Analyze competitor content strategies

Lead Generation

• Profile research: Gather info on prospects
• Connection outreach: Research before connecting
• Content targeting: Find people engaging with relevant topics

---

API Integration

Using the Apify API (JavaScript)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_API_TOKEN' });

const run = await client.actor('localhowl/linkedin-mcp-server').call({
  mode: 'demo',
  tool: 'analyze_voice',
  profileUrl: 'https://www.linkedin.com/in/johnrippy/'
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items[0]);

Using cURL

curl -X POST "https://api.apify.com/v2/acts/localhowl~linkedin-mcp-server/runs?token=YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "demo",
    "tool": "scrape_profile",
    "profileUrl": "https://www.linkedin.com/in/johnrippy/"
  }'

---

Voice Profile Output

When you run analyze_voice, you get:

{
  "voiceProfileId": "johnrippy-voice",
  "profileUrl": "https://www.linkedin.com/in/johnrippy/",
  "postsAnalyzed": 20,
  "profile": {
    "tone": "Professional but conversational",
    "formality": 0.6,
    "hookStyles": [
      "Contrarian statements",
      "Personal stories",
      "Numbers/statistics"
    ],
    "ctaStyles": [
      "Questions to audience",
      "Call to connect",
      "Share your experience"
    ],
    "signaturePhrases": [
      "Here's the thing",
      "Unpopular opinion",
      "Let me explain"
    ],
    "topicCategories": [
      "Career advice",
      "Productivity",
      "LinkedIn growth"
    ],
    "postLength": {
      "average": 850,
      "min": 400,
      "max": 1200
    },
    "hashtagUsage": {
      "average": 3,
      "common": ["#productivity", "#career", "#growth"]
    }
  }
}

---

Limitations

• LinkedIn TOS: Use responsibly, respect rate limits
• Cookie expiry: LinkedIn cookies expire; refresh periodically
• AI accuracy: Generated content should be reviewed before posting
• Profile access: Can only scrape public profiles or connections

---

Support

• Email: john@johnrippy.link
• GitHub: Report issues on the repository

---

Built by John Rippy | johnrippy.link

---

Keywords

linkedin mcp server, model context protocol, claude linkedin, ai linkedin integration, linkedin profile scraper, linkedin voice trainer, linkedin post generator, mcp tools, claude desktop, ai assistant tools, linkedin automation, content generation, ghostwriting ai, personal branding ai