Can Claude Post on LinkedIn? Yes, Here Is Exactly How

This question comes up because Claude writes a very good LinkedIn post, and then nothing happens. You ask it to publish, and it explains that it cannot. That is not Claude being cautious — it genuinely has no way to reach LinkedIn until you give it one.

Why can’t Claude post to LinkedIn by itself?

Claude is a language model. On its own it can read what you type and write text back. It has no account on LinkedIn, no permission to act as you, and no way to send anything to LinkedIn’s servers. There is no hidden setting to switch on.

So out of the box, here is the honest split:

Everything in the “no” column needs one extra piece: a connector.

What is a connector, in plain English?

A connector is a small service that sits between Claude and another app, and gives Claude a set of buttons it is allowed to press. In Claude these are built on MCP — the Model Context Protocol, an open standard for plugging AI assistants into outside tools. Claude calls them custom connectors; you add one in Settings.

The connector is the part that actually holds permission to touch your LinkedIn account. Claude never gets your LinkedIn login, and never sees your access key. It just asks the connector to do a specific job, like “publish this text”, and the connector does it.

Kalovio is one of these connectors, built specifically for posting to LinkedIn. It is free while in public beta, no card required, and it works on Claude’s free plan.

How does the LinkedIn permission actually work?

This is the part worth understanding, because it is what separates a legitimate tool from a risky one.

When you connect, LinkedIn shows you its own consent screen — on linkedin.com, in your browser. You are granting two things: Sign In with LinkedIn, which confirms who you are, and Share on LinkedIn, which carries the permission called w_member_social. That permission is what allows an app to publish a post on your behalf. Both are free, self-serve LinkedIn products open to any developer.

That grant produces an access key. In Kalovio’s case that key is encrypted with AES-256-GCM (a standard, strong method of scrambling data so it is unreadable without the right key) and stored on the server. It is never handed to Claude, never written into a chat, and never printed in a tool result. You can revoke it any time from your LinkedIn settings.

What can Claude do on LinkedIn once a connector is added?

With Kalovio connected, the things Claude could only describe before become things it can do. In plain conversation, you can ask it to:

• Publish a post now — text up to LinkedIn’s 3,000-character limit.
• Schedule a post for a specific date and time, in your own timezone.
• Fill a weekly queue — set your posting slots once, then drop posts into the next free one.
• Attach one image, or 2 to 9 images as a swipeable carousel.
• Post a PDF carousel — a document post, the format that tends to hold attention in the feed.
• Tag a company page so the mention is a real clickable link, not just text.
• Track link clicks using short links, so you can see which post actually sent traffic.
• Keep drafts, save a voice profile so posts sound like you, and delete a post it published.

The caption is optional when you attach a file, so an image-only or PDF-only post is fine. Files are capped at 20 MB each, 100 files per account, kept for 30 days.

Does Claude post without asking me first?

No, and this is deliberate. Publishing, scheduling, queueing and deleting are all two-step. The first call does not publish anything — it returns a preview of the exact post, plus a signed token that covers that exact content: the words, the image, the publish time, the mentions, the links. Nothing reaches LinkedIn until a second call confirms it, and the token still matches.

Change one character and the token no longer matches, so the confirmation fails. That means what you approved is what goes out. It also blocks a real risk with AI tools: text hidden inside a web page or a document telling the model to post something. Without your explicit approval, that instruction goes nowhere.

What still won’t work, even with a connector?

These are LinkedIn’s limits, not the connector’s, and no tool built on the free self-serve API can get around them:

• Claude cannot read your old posts. The permission for that (r_member_social) is closed to self-serve apps. It can only see posts it published itself.
• No engagement analytics. Impressions, follower growth and reactions need LinkedIn’s partner-gated Community Management API, which requires a registered company and app review. Kalovio has link-click tracking only — useful, but not the same thing.
• No DMs, no connection requests, no lead scraping. Not supported, and not planned.
• Access expires. LinkedIn member tokens last about 60 days and self-serve apps get no automatic renewal, so you reconnect every couple of months. Kalovio gives you a fix-it link when that happens.
• You cannot hand Claude an image from the chat. Claude has no way to forward a chat attachment to a connector. You upload through Kalovio’s upload panel or its browser link instead, then refer to the image by name.
• Tagging a company needs its numeric page ID. LinkedIn gives self-serve apps no name-to-ID lookup, so Kalovio keeps a shared directory of companies it has already seen.

Is this allowed under LinkedIn’s rules?

Worth being straight about. LinkedIn’s API Terms of Use, section 3.1(26), restricts using the APIs “to automate posting on the LinkedIn Services”, and there is no written exception for member-approved posting. So nobody can honestly promise you zero risk, and you should be sceptical of any tool that does.

The defensible position is narrower and more specific: the permission is open and self-serve, you grant it yourself through LinkedIn’s own consent screen, you approve the exact words of every single post before it goes out, nothing is scraped, and you can revoke access at any moment. That is a person publishing with help, not a bot running unattended.

How do I set it up?

It is a short job — add the connector in Claude’s settings, sign in with LinkedIn once, and you are posting. The step-by-step version, with the exact URL and screenshots of each screen, is in the guide to connecting LinkedIn to Claude.

Then the answer to the original question stops being theoretical. You draft in the same chat where you think, you read the preview, you say yes, and the post is live.