"LinkedIn MCP" is a fast-rising search, so here's the honest picture: LinkedIn has no broad official MCP server, and the community options that scrape your account live in a grey area with real account-risk. This guide covers what exists and how to use it carefully.
Caveat first: LinkedIn's terms restrict automated access and scraping. Community servers that log in as you can get your account flagged or banned. Proceed knowingly, and prefer official APIs where your use case allows.
The landscape
- Community scraper servers (e.g. self-hosted servers that drive a logged-in session) can read profiles, search and pull post data. Powerful, but they use unofficial access — that's the ToS and ban risk.
- Official LinkedIn APIs (Marketing, Share, Sign In) are sanctioned but narrow and require an approved developer app — better for posting/marketing use cases than for broad scraping.
- Bridges (Composio, Pipedream) expose LinkedIn actions through their managed integrations within LinkedIn's allowed scopes.
Typical community setup
Most run via Docker or uv and need your LinkedIn session/cookie:
{
"mcpServers": {
"linkedin": {
"command": "uvx",
"args": ["<community-linkedin-mcp-package>"],
"env": { "LINKEDIN_COOKIE": "your_li_at_cookie" }
}
}
}
Swap in the exact package from the repo you vet. Restart the client; see MCP config file location.
Safer path
For posting and analytics, use an official LinkedIn developer app + a bridge like Composio that stays within sanctioned scopes — lower ban risk than a cookie-based scraper.
Use responsibly
Read your session cookie like a password, never share it, and understand the account risk before automating LinkedIn. Review any community server's code first — see how to vet an MCP server and MCP security best practices.
Going further
Browse the communication category or curated loadouts.