What it does
This MCP server connects Claude to a LinkedIn account via a Chromium browser session, exposing tools to retrieve and interact with profiles, companies, jobs, and messaging. It supports 16 operations spanning profile lookup (self or others), company research, job search, people search, messaging, and sidebar recommendations. All authentication happens through a standard LinkedIn login flow; the server manages a persistent browser session and caches Chromium under ~/.linkedin-mcp/patchright-browsers. Tool execution is sequential with no built-in rate limiting, and the server can send messages (with confirmation), search conversations, retrieve feed posts, and access detailed job information.
Who it's for
Recruiters building candidate pipelines, job seekers researching opportunities, sales development reps finding prospects, and researchers analyzing professional networks—anyone integrating LinkedIn data into AI workflows.
Common use cases
- Search for job postings in specific locations and retrieve full job details
- Find and retrieve profile information for people matching location and company filters
- Research company details, employee lists, and recent posts
- Monitor and search LinkedIn messages without manual inbox navigation
- Build connection-finding workflows by retrieving recommendations from profile sidebars
Setup pitfalls
- LinkedIn's Terms of Service prohibit automated tools; while the server uses a real browser (not API exploitation), high-volume automation risks account warnings or restrictions
- Requires browser authentication on first use, with the login window opening automatically when any tool first runs
- No built-in rate limiting—agents can trigger rapid requests and activate LinkedIn's abuse detection; design prompts to avoid bulk operations
- Writes browser cache and state files to
~/.linkedin-mcp/patchright-browsers; ensure adequate disk space and file permissions