linkedin-scraper-mcp

What it does

Exposes LinkedIn profiles, companies, jobs, and messaging as MCP tools for Claude. Search for people and companies, fetch detailed profile sections (experience, education, skills, posts), list company employees, search jobs, read conversations, and send messages. Automates browser-based authentication and maintains sessions using Patchright Chromium.

Who it's for

Recruiters automating candidate research, sales development reps doing account prospecting, and business intelligence teams gathering company data at scale.

Common use cases

• Search for job postings by title and location, then fetch detailed requirements and company info.
• Extract employee lists from a company's LinkedIn page and filter by keyword.
• Retrieve detailed profile sections (experience, education, certifications) for a candidate.
• Read and search messaging conversations to automate outreach workflows.
• Search for people by keywords, location, and current company affiliation.

Setup pitfalls

• Browser authentication required on first use—the server opens a browser window to log into LinkedIn. Early tool calls before login completes return errors; either wait for auth or run uvx linkedin-scraper-mcp --login explicitly first.
• Requires filesystem write access to cache the Chromium browser under ~/.linkedin-mcp/patchright-browsers. Restrict this directory appropriately in sandboxed environments.
• Default HTTP timeout may be insufficient for large data fetches. Set UV_HTTP_TIMEOUT=300 or higher in environment configuration.
• High risk class due to LinkedIn's terms of service—use responsibly and ensure compliance with LinkedIn's acceptable use policies and local laws.