/ 目錄 / 演練場 / linkedin-mcp-server
← 全部伺服器 ↗ GitHub
● 社群 stickerdaniel ⚡ 即開即用

linkedin-mcp-server

作者 stickerdaniel · stickerdaniel/linkedin-mcp-server

Give AI assistants access to LinkedIn — browse profiles, search jobs, read messages, and send connection requests.

linkedin-mcp-server is an open-source MCP server that enables AI assistants to interact with LinkedIn via browser automation. It provides 13 tools for accessing profiles, company pages, job listings, messaging, people search, and connection management. Uses Patchright (managed Chromium) for browser-based access with automatic login handling.

▶ 觀看演示 📦 安裝 💡 使用場景

為什麼要用

核心特性

即時演示

實際使用效果

linkedin-mcp-server-2.replay ▶ 就緒
0/0

安裝

選擇你的客戶端

~/Library/Application Support/Claude/claude_desktop_config.json  · Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "linkedin-mcp-server-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/stickerdaniel/linkedin-mcp-server"
      ]
    }
  }
}

開啟 Claude Desktop → Settings → Developer → Edit Config。儲存後重啟應用。

~/.cursor/mcp.json · .cursor/mcp.json
{
  "mcpServers": {
    "linkedin-mcp-server-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/stickerdaniel/linkedin-mcp-server"
      ]
    }
  }
}

Cursor 使用與 Claude Desktop 相同的 mcpServers 格式。專案級設定優先於全域。

VS Code → Cline → MCP Servers → Edit
{
  "mcpServers": {
    "linkedin-mcp-server-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/stickerdaniel/linkedin-mcp-server"
      ]
    }
  }
}

點擊 Cline 側欄中的 MCP Servers 圖示，然後選 "Edit Configuration"。

~/.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "linkedin-mcp-server-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/stickerdaniel/linkedin-mcp-server"
      ]
    }
  }
}

格式與 Claude Desktop 相同。重啟 Windsurf 生效。

~/.continue/config.json
{
  "mcpServers": [
    {
      "name": "linkedin-mcp-server-2",
      "command": "TODO",
      "args": [
        "See README: https://github.com/stickerdaniel/linkedin-mcp-server"
      ]
    }
  ]
}

Continue 使用伺服器物件陣列，而非映射。

~/.config/zed/settings.json
{
  "context_servers": {
    "linkedin-mcp-server-2": {
      "command": {
        "path": "TODO",
        "args": [
          "See README: https://github.com/stickerdaniel/linkedin-mcp-server"
        ]
      }
    }
  }
}

加入 context_servers。Zed 儲存後熱重載。

claude mcp add linkedin-mcp-server-2 -- TODO 'See README: https://github.com/stickerdaniel/linkedin-mcp-server'

一行命令搞定。用 claude mcp list 驗證，claude mcp remove 移除。

使用場景

實戰用法： linkedin-mcp-server

How to research companies and job openings with LinkedIn MCP

👤 Job seekers and recruiters ⏱ ~15 min beginner

何時使用： You want to research companies and find relevant job openings without manual browsing.

前置條件
  • linkedin-mcp-server installed — uvx linkedin-scraper-mcp@latest or download .mcpb from releases
  • LinkedIn account — Sign in when the browser login prompt appears
步驟
  1. Search for jobs
    Search LinkedIn for 'senior backend engineer' jobs in San Francisco. Show me the top 10 with company, salary range, and requirements.✓ 已複製
    → Formatted job listings with key details
  2. Research a company
    Get the company profile for the most interesting result. What are their recent posts about?✓ 已複製
    → Company overview with recent activity

結果： A shortlist of relevant jobs with company context.

注意事項
  • LinkedIn rate limits on too many profile views — Space out requests and avoid viewing dozens of profiles in rapid succession
搭配使用： filesystem

Research and connect with people in your industry

👤 Professionals building their network ⏱ ~20 min beginner

何時使用： You want to find and connect with relevant professionals in your field.

前置條件
  • LinkedIn account authenticated — Complete the browser login on first use
步驟
  1. Search for people
    Search LinkedIn for ML engineers at top AI companies in the Bay Area. Show their titles and companies.✓ 已複製
    → List of matching professionals
  2. Send a personalized connection request
    Send a connection request to the first person with a note mentioning our shared interest in transformer architectures.✓ 已複製
    → Connection request sent confirmation

結果： Targeted connection requests sent with personalized messages.

注意事項
  • Generic connection messages get ignored — Always reference something specific from their profile

組合

與其他 MCP 搭配，撬動十倍槓桿

linkedin-mcp-server-2 + filesystem

Save job search results and company profiles to organized files

Search for product manager jobs in NYC and save the top 20 results to ~/job-search/pm-nyc.md.✓ 已複製

工具

此 MCP 暴露的能力

工具輸入參數何時呼叫成本
get_person_profile url_or_name: str, sections?: str[] Get a LinkedIn profile with selected sections 0
search_jobs keywords: str, location?: str Search for job listings 0
search_people keywords: str, location?: str Find professionals by keywords 0
get_inbox none List recent LinkedIn message conversations 0
get_company_profile url_or_name: str Get company information from LinkedIn 0
connect_with_person profile_url: str, note?: str Send a connection request 0

成本與限制

運行它的成本

API 配額
Subject to LinkedIn's anti-automation measures — keep requests moderate
每次呼叫 Token 數
500–2000 tokens per profile/search
費用
Free
提示
Be selective with profile views — LinkedIn tracks and may restrict accounts that view too many profiles.

安全

權限、密鑰、影響範圍

憑證儲存： Uses your LinkedIn browser session. Authentication happens via Chromium browser login.
資料出站： All traffic goes through Chromium to linkedin.com

故障排查

常見錯誤與修復

Login prompt keeps appearing

Complete the LinkedIn login in the Chromium browser. For Docker, authenticate via uvx first.

驗證： Check that you can manually access linkedin.com in the automated browser
Profile returns empty data

The profile may be outside your network. LinkedIn restricts data visibility based on connection degree.

驗證： Try viewing the profile manually in a browser
Rate limited by LinkedIn

Reduce request frequency. LinkedIn may temporarily restrict access after too many automated requests.

驗證： Wait 30 minutes and try a single request

替代方案

linkedin-mcp-server 對比其他方案

替代方案何時用它替代權衡
fetch MCPYou only need to read public LinkedIn data visible without loginVery limited data — most LinkedIn content requires authentication

更多

資源

📖 閱讀 GitHub 上的官方 README

🐙 查看未解決的 issue

🔍 瀏覽全部 400+ MCP 伺服器和 Skills