LinkedIn MCP Server

What is LinkedIn MCP Server?

An MCP server that enables AI assistants like Claude to read LinkedIn data through your own logged-in browser session. It allows access to profiles, companies, job searches, and messaging. This is an independent community project not affiliated with LinkedIn.

How to use LinkedIn MCP Server?

Install uv, then configure your MCP client with the command uvx mcp-server-linkedin@latest and set the environment variable UV_HTTP_TIMEOUT to 300. On the first tool call, a browser window opens for LinkedIn login. Alternatively, use the --login flag to set up a persistent session or --import-from-browser to reuse an existing browser session. Command-line options allow customising browser visibility, timeouts, transport, and profile location.

Key features of LinkedIn MCP Server

• Read profiles with section selection (experience, education, skills, etc.)
• Search for people, companies, and jobs with filters
• Send connection requests and messages
• View company profiles, posts, and employees
• Access messaging inbox and conversations
• Search conversations by keyword
• Get your own profile and feed posts
• Import session from common browsers (Chrome, Brave, Edge, etc.)

Use cases of LinkedIn MCP Server

• Automate LinkedIn profile research during recruiting or sales prospecting.
• Monitor job listings matching specific criteria via AI assistant.
• Gather company intelligence (posts, employees, profiles) for market analysis.
• Manage LinkedIn networking (connection requests, messages) programmatically.
• Analyze your own feed or inbox through AI queries.

FAQ from LinkedIn MCP Server

What dependencies are required?

Python and uv are needed. The server uses Patchright Chromium (downloaded automatically) and requires a network connection to LinkedIn.

How does authentication work?

Authentication is handled via your own browser session. The server opens a browser for you to log in to LinkedIn on the first tool call, or you can pre-configure login with --login or import cookies from an existing browser session using --import-from-browser.

Where does my data live?

All data is retrieved from LinkedIn through your session and processed locally. No data is stored externally by the server; session cookies are saved locally in ~/.linkedin-mcp/profile/.

What transport modes are supported?

The default transport is stdio. Streamable HTTP is also available for web-based MCP servers. You can set transport via --transport option or through environment variables.

Are there any known limitations?

LinkedIn's page structure changes frequently