Is There a LinkedIn Connector for Claude?

This question gets asked because Claude clearly connects to things — you can see a list of connectors in its settings — and LinkedIn is not in it. So people reasonably wonder whether they missed a switch somewhere. They did not. The answer has two halves, and only the second one is the useful part.

Does Claude have a built-in LinkedIn connector?

No. Anthropic ships a set of connectors that appear in Claude by default, and LinkedIn is not among them. LinkedIn does not publish one either — there is no official LinkedIn connector or MCP server from LinkedIn itself, for Claude or for anything else.

So if you are looking for a first-party toggle labelled “LinkedIn”, it does not exist and you can stop hunting for it.

So how does LinkedIn work in Claude at all?

Because Claude does not limit you to its own list. It supports custom connectors: you paste a web address, Claude connects to whatever is at that address, and the tools there appear in your chat.

Under the hood, a connector is an MCP server. MCP stands for Model Context Protocol — an open standard for letting an AI assistant talk to an outside service through a defined set of tools. A remote MCP server is just that standard, running on a website you connect to. In Claude’s interface these are called connectors; in the wider ecosystem they are called MCP servers. Same thing.

That is the real answer to the question. A LinkedIn connector for Claude exists in the same way a LinkedIn app exists for your phone: someone other than the platform built it, and you add it yourself.

What LinkedIn connectors actually exist?

They split cleanly into two groups, and the split matters more than any feature list.

Group one reads LinkedIn. These pull profiles, company pages, job listings or search results. The most visible example is stickerdaniel/linkedin-mcp-server, the top Google result for “linkedin mcp”. It is a scraper — it drives a browser using your logged-in session — and its own README says it “does not publish or create posts on your behalf”, and warns that accounts “can be restricted or banned”. Apify and Bright Data also offer LinkedIn MCP endpoints, again for data extraction rather than posting.

Group two writes to LinkedIn. This group is much smaller, because publishing needs LinkedIn’s official permission system rather than a borrowed browser session. Zapier’s LinkedIn MCP exposes only about two actions. Taplio’s own page says its MCP is “text-only... you cannot upload images, video, or PDF carousels”. ContentIn’s LinkedIn MCP server does not exist yet — it is currently in build, with a waitlist.

Which one lets Claude actually post?

If posting is what you came for, Kalovio is the free hosted one. You add it to Claude as a custom connector at https://kalovio.com/mcp, sign in with LinkedIn once, and then write, publish, schedule and queue posts in plain conversation.

The part worth understanding is how it signs in, because that is what separates the two groups above. Kalovio uses LinkedIn’s own OAuth flow — the “Continue with LinkedIn” screen — with the w_member_social permission, which allows publishing on your behalf and nothing else. It never asks for a cookie, never automates a browser, and never scrapes. Your LinkedIn token is encrypted and stays on the backend; it is never handed to Claude and never appears in a tool result.

It also works on Claude’s free plan, which is not true of the ChatGPT equivalent — there, connectors of this kind need Developer Mode, which is on paid plans and web only. If you are comparing the two, see can Claude post on LinkedIn.

How do I add it?

Four steps, once:

• In Claude, open Settings → Connectors → Add custom connector.
• Paste the connector’s URL. For Kalovio that is https://kalovio.com/mcp.
• Claude opens a sign-in page. Approve it, then approve the LinkedIn screen that follows. That second screen is on linkedin.com and is where you can see exactly what is being requested.
• Back in a chat, ask it to draft something. The connector’s tools show up automatically.

There is a longer walkthrough with screenshots-worth of detail at connect LinkedIn to Claude, including what to do if the connector does not appear in the list.

What can a LinkedIn connector for Claude do?

With Kalovio specifically: write and publish a post, save and edit drafts, schedule a post for a future time, build a posting queue, attach a single image or two to nine images as a swipeable set, post a PDF carousel, mention companies, set a voice profile so drafts sound like you, track clicks on links you share, and delete a post it published. Posts can be up to 3,000 characters. A caption is optional when an image or PDF is attached.

Every publish, schedule, queue and delete is two-step. The first call returns a preview plus a signed token covering the exact content; nothing happens until you confirm. That gate exists so hidden instructions in a web page or document the model read cannot quietly post something.

What can it not do?

Worth knowing before you connect anything, because these limits come from LinkedIn, not from the tool:

• It cannot read posts it did not create. The permission for that, r_member_social, is closed to self-serve apps. See why it cannot read your old posts.
• No engagement analytics. Likes, comments and impressions need the partner-gated Community Management API. Link-click tracking is available, and it is not the same thing.
• Sign-in expires after about 60 days. Self-serve apps get no refresh token, so you reconnect. It is one click.
• Company tagging needs the page’s numeric id. There is no name lookup for self-serve apps, so known companies resolve from a shared directory and new ones need the number once.
• Multi-image and PDF posts use LinkedIn’s newer versioned interface, which may be product-gated for a self-serve app. A refusal is reported clearly rather than failing silently — try it on your own account and do not assume it works for everyone.
• PDF posts cannot be scheduled yet — immediate publish only.
• No messages, no connection requests, no lead scraping.
• Neither Claude nor ChatGPT can hand a chat attachment to a connector. You upload images through Kalovio’s own upload panel or link instead.

What should I check before adding any LinkedIn connector?

Anything you add gets to act inside your chats, so it is worth thirty seconds of checking. Four questions do most of the work:

• Does connecting send you to linkedin.com? If setup asks you to paste a session cookie, hand over a password, or install a browser extension that reads your session, it is not using OAuth. That single test sorts the two groups above faster than any feature comparison.
• Can it name the permission it uses? A tool posting through the official route will say w_member_social plainly. Vague phrases like “official integration” with no named permission are just an assertion.
• Can you revoke it from LinkedIn’s own settings? An OAuth grant is listed there and switches off in one click. A borrowed cookie is not listed anywhere, so the only way to kill it is to change your password or sign out of every device.
• Does anything publish without your approval? An assistant reads web pages and documents, and those can carry hidden instructions. A connector that posts on a single tool call has no defence against that. One that shows you a preview and waits for a confirmation does.

If you want that reasoning in more depth, is a LinkedIn MCP server safe covers the cookie-versus-OAuth split properly, and the comparison page lines the options up side by side.

Is adding a LinkedIn connector allowed?

The short version

Yes, there is a LinkedIn connector for Claude — several, in fact. None of them come from Anthropic or LinkedIn. You add one yourself as a custom connector, and the first thing to check is whether it reads or writes, because most of the popular ones only read.