Is There an Official LinkedIn MCP Server? (Short Answer: No)

If you searched for this, you probably found a dozen GitHub repositories and a few paid tools, all calling themselves “the LinkedIn MCP server”, and none of them made it clear whether LinkedIn itself was behind any of it. Here is the straight answer, and the distinction that actually matters when you pick one.

Does LinkedIn make an MCP server?

No. As of August 2026, LinkedIn (and its parent, Microsoft) does not ship, host, or endorse an MCP server for LinkedIn. There is no mcp.linkedin.com, no entry in LinkedIn’s developer product list, and no official announcement.

MCP — the Model Context Protocol — is an open standard for connecting AI assistants to outside tools. It was published by Anthropic and is now used by Claude, ChatGPT and others. Anyone can build an MCP server for any service. That is exactly why so many LinkedIn ones exist: the protocol is open, LinkedIn’s participation is not required to build one.

So every result you find is a third-party project. The real question is not “is this official?” — none of them are. The question is what is it built on top of?

The distinction that actually matters: official product vs official API

These two phrases sound the same and mean very different things.

LinkedIn does publish a genuinely official, free, self-serve API. Two products matter here:

• Sign In with LinkedIn (OIDC) — lets an app confirm who you are, with your consent.
• Share on LinkedIn — grants the w_member_social permission, which lets an app publish posts on your behalf, with your approval.

Both are open to any developer, with no payment and no partner status. That means a third-party MCP server can be built entirely on sanctioned rails — no scraping, no borrowed cookies — even though LinkedIn itself does not publish the server.

What actually exists today

Search results for “LinkedIn MCP” fall into three groups.

1. Open-source scrapers (the most common)

These read LinkedIn through your logged-in browser session or a copied cookie. They can fetch profiles, companies and job listings. The most popular one is upfront about its limits — its own README states that it “does not publish or create posts on your behalf” and warns that accounts “can be restricted or banned”.

So if your goal is to write and publish posts, the single most popular LinkedIn MCP server cannot do it, by its author’s own admission.

2. Data-extraction platforms

Services like Apify and Bright Data offer LinkedIn MCP endpoints aimed at bulk data collection. Same story: built for reading and extracting, not for posting as you.

3. Posting tools built on the official API

A smaller group uses w_member_social to publish. These can genuinely post for you. Most are paid, on a trial, or still on a waitlist. Kalovio is in this group and is free while in beta.

How to check any tool’s claim in 60 seconds

“Official API, no scraping” is easy to write on a landing page and rarely proven. Ask for specifics. A tool genuinely built on LinkedIn’s API can name, without hesitation:

• The permission scope it requests — for posting, that is w_member_social.
• The endpoint it posts to — POST /v2/ugcPosts, or LinkedIn’s versioned /rest/posts for multi-image and document posts.
• The sign-in method — a real OAuth consent screen on linkedin.com, not a box asking you to paste a cookie.

That last one is the fastest test of all. If a tool asks you to paste a cookie or your LinkedIn password, it is not using the official API. A real OAuth flow sends you to LinkedIn’s own page, shows you exactly what you are granting, and can be revoked any time from your LinkedIn settings.

What no LinkedIn MCP server can do

Because there is no official server, everything runs on the same public rails — which have real limits nobody can engineer around:

• Reading your old posts is not possible through the self-serve API. The permission that would allow it (r_member_social) is closed. A tool can only see posts it created itself.
• Access expires. LinkedIn member tokens last about 60 days, and self-serve apps get no refresh token, so you will reconnect periodically. Any tool claiming permanent access is either scraping or wrong.
• Real engagement analytics are gated. Impressions and follower stats need LinkedIn’s partner-only Community Management API, which requires a registered company and app review.
• Tagging a company page needs its numeric ID. LinkedIn gives self-serve apps no way to look up a company by name.

So what should you actually use?

It depends entirely on the job:

• Reading profiles, companies or jobs in bulk? The scraper repos and data platforms do that. Understand the account risk you are accepting, because their own documentation warns about it.
• Writing and publishing your own posts? You need something built on w_member_social. That is what Kalovio does — you connect your own LinkedIn account through LinkedIn’s own consent screen, and you approve every post before it goes out.

There is no official LinkedIn MCP server to wait for. There is a legitimate official API, and it is worth checking which side of that line your tool sits on.