ContactGraphLinkedIn MCP server
The LinkedIn MCP server built for job search — not posting.
Developers are building LinkedIn MCP servers to let AI tools read profiles and post content. ContactGraph’s MCP server does something different: it connects Claude to your exported LinkedIn connections so you can find warm paths into companies that are hiring — without scraping or account risk.
Last updated June 2026
The problem
Why existing LinkedIn MCP servers don’t help with job search
Search interest in “linkedin mcp” and “linkedin mcp server” has surged as professionals try to connect AI assistants to their LinkedIn data. The most popular open-source LinkedIn MCP servers (with thousands of GitHub stars) focus on two use cases: posting content and scraping profiles through a browser session.
Neither helps with the use case most job seekers actually need: searching your own network to find who you know at a company with open roles. Scraping MCP servers can technically search LinkedIn, but they operate outside LinkedIn’s terms and risk your account. Posting MCP servers use the official API but can’t read your connection list at all.
For job search, what you need is an MCP server that knows your contacts, knows which companies are hiring, and can match the two. That requires owning your network data — not renting access to LinkedIn’s platform.
Available approaches
How people try to connect AI to LinkedIn
Profile-scraping MCP servers
e.g. stickerdaniel/linkedin-mcp-server — reads LinkedIn through browser automation.
Pros
- +Can search people, companies, and jobs on LinkedIn
- +One-click Claude Desktop install (.mcpb bundle)
- +Active open-source community
Cons
- –Scraping violates LinkedIn TOS
- –Account suspension risk
- –Doesn’t know YOUR connections — searches all of LinkedIn
- –No warm-intro path analysis
OAuth posting MCP servers
e.g. gacabartosz/linkedin-mcp-server — posts, schedules, and manages content via official API.
Pros
- +Compliant with LinkedIn’s posting API
- +Rich posting features (images, scheduling, templates)
- +24+ tools for content management
Cons
- –Zero job-search capabilities
- –Cannot access your connection list
- –Requires LinkedIn Developer app setup
- –Focused on content marketing, not networking
ContactGraph MCP server
Queries your exported LinkedIn connections merged with phone contacts — matched to live job listings.
Pros
- +Built specifically for warm-path job search
- +Shows open roles at companies where you know someone
- +Enriches contacts with current employer and role
- +Open source MCP server — free, no API keys required
- +No scraping, no account risk
Cons
- –Requires one-time LinkedIn data export
- –Network data is a snapshot until you re-export
- –Not designed for posting to LinkedIn
The solution
Set up ContactGraph as your LinkedIn MCP server for job search
ContactGraph’s MCP server exposes your professional network as structured tools that Claude can query. Unlike scraping servers, it works entirely from data you own.
- 1
Export and upload your LinkedIn connections
Download Connections.csv from LinkedIn’s data export and upload it to ContactGraph along with your phone contacts. ContactGraph enriches each contact with current employer, title, and industry.
- 2
Connect the MCP server to Claude
Add ContactGraph as a custom MCP connector in Claude Desktop, or run the server locally via Claude Code. The server exposes tools for searching contacts, finding companies, and matching jobs.
- 3
Search for warm paths to open roles
Ask Claude: “Which companies in my network are hiring for engineering roles?” or “Who can introduce me to someone at Notion?” Claude queries your graph and returns actionable warm-intro paths.
- 4
Act on warm introductions
Use Claude to draft personalized outreach based on your actual relationship context — not generic templates. A warm intro through someone you know beats a cold application every time.
FAQ
Frequently asked questions
- What is a LinkedIn MCP server?
- A LinkedIn MCP server is a Model Context Protocol server that gives AI assistants like Claude access to LinkedIn data and actions. MCP is an open standard by Anthropic that lets AI tools call external APIs through a standardized interface.
- Which LinkedIn MCP server is best for job search?
- For job search, ContactGraph’s MCP server is the best fit because it searches your actual connections and matches them to open job listings. Scraping-based servers search all of LinkedIn (not your network specifically) and risk account suspension. Posting servers can’t search contacts at all.
- Can I use a LinkedIn MCP server with Claude Desktop?
- Yes. Both scraping MCP servers and ContactGraph work with Claude Desktop via custom connectors or local MCP configuration. ContactGraph also works with Claude Code, Cursor, and any MCP-compatible client.
- Is it safe to use a LinkedIn MCP server?
- It depends on the approach. Scraping-based servers violate LinkedIn’s terms and can get your account restricted. ContactGraph uses LinkedIn’s official data export — the same file LinkedIn lets every user download — so there is no platform risk.
- How does ContactGraph find jobs to match with my network?
- ContactGraph aggregates open roles from across the web and matches them to companies where you have a contact in your graph. Every job listing comes with a warm path — someone you know who can make an introduction.
- Do I need to pay for a LinkedIn MCP server?
- ContactGraph is free — open source, nonprofit, no subscription. Many alternatives charge $10–50/month (ConnectSafely, LinkupAPI, Composio) or require you to run fragile scraping infrastructure yourself.
Get started
Export once. Connect any AI. Own your network.
Upload your LinkedIn connections and phone contacts. ContactGraph builds your graph and exposes it to Claude, ChatGPT, and any MCP client — free, open source, and never sells your data.
Nonprofit · Open source · Never shares your data · Free forever