Search for "LinkedIn MCP server" and you will find GitHub projects, vendor pages, MCP directories, and comparison articles. The confusing part is that many of them are solving completely different problems.
Some are built for publishing your own LinkedIn content. Some are built for prospecting and outreach. Some only read data. And some use LinkedIn's official API, while others drive your logged-in browser or use your session.
That difference matters because it changes what the server can do, how safe the setup is, and whether it fits your actual goal. This guide explains the four main types of LinkedIn MCP servers and compares 11 options.
Key takeaways
- LinkedIn MCP tools fall into four broad groups: content platforms, automation hubs, session-based services, and scrapers or outreach tools. There is also LinkedIn's official people-lookup integration in ChatGPT, but it is not an MCP server for posting.
- If your goal is publishing your own content, use a content platform that works through LinkedIn's official API.
- Session-based tools can do more in some cases, but they conflict with LinkedIn's rules.
How a LinkedIn MCP Server Works
The server exposes tools that an assistant such as Claude, ChatGPT, or Cursor can call when you ask for something. Those tools might include actions such as:
- Draft a post
- Schedule a post
- Search profiles
- Read analytics
- Create comments
- Fetch prospect data
MCP was introduced by Anthropic in late 2024 and is now supported by most major assistants. A simple way to think about it is as a universal connector: instead of every AI assistant needing a different custom integration for every app, MCP gives them a standard way to connect to tools.
What a LinkedIn MCP server can actually do depends on two things:
- Which tools the server exposes
- How that server reaches LinkedIn
Does LinkedIn Have an Official MCP Server?
No. Here is what we found on October 2, 2026:
- The official MCP registry returned 38 distinct servers for "linkedin."
- None was published by LinkedIn or Microsoft.
- One listing,
ai.com.mcp/linkedin, looks official at first glance, but its own description credits a third-party service. - LinkedIn does have its own ChatGPT plugin for looking up people, but it cannot publish posts.
What LinkedIn does provide is an official API. Approved applications can use that API to publish on a user's behalf after the user signs in and approves access. Any MCP server that publishes safely through LinkedIn is therefore built on top of an app that has the necessary LinkedIn API access.
The Four Kinds of LinkedIn MCP Servers
The easiest way to compare LinkedIn MCP tools is to ask two questions:
- Does the tool reach LinkedIn through an approved app, or through your own logged-in session?
- Is the tool mainly for publishing content, or for data and outreach?
- FeedBoss, ContentIn, Postbeam: post through LinkedIn's official API, by their own description
- Taplio, Supergrow, Buffer: post through your account with them
- Composio, Zapier: raw posting actions after a LinkedIn sign-in
Best fit for publishing your own content.
- Unipile: connects with your LinkedIn username and password or session cookie, then publishes and comments
- Developer tools first; not built for creators
Works through the account itself, which LinkedIn's terms restrict.
- LinkedIn's own ChatGPT plugin: looks up people
- No LinkedIn-published server in the official MCP registry
The only first-party option we found, and it cannot post.
- stickerdaniel/linkedin-mcp-server: reads profiles, jobs, and posts through your logged-in browser
- gtm-api: search, connect, and message from an account in a cloud browser
- Apify LinkedIn actors: data extraction; the one we checked uses your cookies
Built for prospecting and data, not for your own posts.
1. Content Platforms
These are LinkedIn content tools that added MCP support. FeedBoss, ContentIn, and Postbeam each describe publishing through LinkedIn's official API. Taplio, Supergrow, and Buffer also publish through the user's account with their platforms.
Since they can all post, the differences that matter are things such as:
- Whether they can write in your voice
- Whether they review the draft before publishing
- Whether they support images
- How they are installed
- Whether they include a content calendar or analytics
2. Automation Hubs
Composio and Zapier expose lower-level LinkedIn actions after you connect a LinkedIn account. For example, they can create posts and comments.
They are flexible building blocks, but they do not include a built-in voice system, review workflow, or complete scheduling flow unless you build one. Zapier's post action publishes immediately.
3. Session-Based Servers
These tools work directly through your LinkedIn account session. That can mean your logged-in browser, your session cookie, or your LinkedIn username and password. Unipile, for example, connects through your password or cookie and can then publish posts and comments.
This approach can expose more actions than the official API, but it comes with an important tradeoff, covered in the safety section below.
4. Scrapers and Outreach Tools
These tools are mainly built for reading LinkedIn data, searching, prospecting, or sending outreach. The most popular open-source LinkedIn MCP server we found was stickerdaniel/linkedin-mcp-server, with about 3,700 GitHub stars. It reads profiles, companies, jobs, and posts through your logged-in browser and can also handle messages, but we found no tool for publishing your own posts.
Other examples include:
gtm-api, which runs accounts in a cloud browser for search, connections, and messages.- Apify LinkedIn actors, which extract LinkedIn data. The actor we checked uses your cookies.
LinkedIn MCP Servers Compared
The table below uses each vendor's own description as of October 2, 2026. "Not stated" means the vendor's MCP documentation does not say. It does not necessarily mean the feature is missing.
| Server | How it reaches LinkedIn | What it does with posts | Writes in your voice | Cost to use the MCP |
|---|---|---|---|---|
| FeedBoss | Official LinkedIn API, after you sign in | Drafts, checks for AI slop, makes on-brand images, reads your analytics, schedules after you approve a preview | Yes, from your own posts and knowledge base | From $29 a month, with 1,000 credits (about 200 posts) |
| Taplio | Your Taplio account; Taplio says it does not use your LinkedIn password or cookies | Publishes or schedules; Taplio says publishing needs your confirmation | Not stated | From $39 a month; AI writing needs the $69 plan, per Taplio |
| ContentIn | LinkedIn API, per ContentIn | Publishes or schedules after showing you the text | Yes, per ContentIn | ContentIn Pro plan |
| Supergrow | Not stated | Queues posts for approval; scores hooks, clarity, and calls to action | Yes, per Supergrow | Any paid Supergrow plan |
| Postbeam | LinkedIn's official partner APIs, per Postbeam | Drafts and schedules; nothing publishes without approval | Yes, per Postbeam | From $39 a month after a trial |
| Buffer | Not stated | Drafts and schedules into your Buffer queue for review | Not stated | Every Buffer plan, including Free |
| Composio | LinkedIn sign-in through Composio | Raw tools to create posts and comments; you build the workflow | No | Free Hobby plan available |
| Zapier | LinkedIn sign-in through Zapier | Publishes immediately; no built-in scheduling | No | Every Zapier plan; each call uses tasks |
| Unipile | Your LinkedIn username and password, or session cookie | Publishes posts and comments; built for developers | No | 49 EUR a month for up to 10 accounts |
| stickerdaniel/linkedin-mcp-server | Your logged-in browser session | Reads profiles, jobs, and posts; messages; no publishing tool found | No | Free, open source |
| gtm-api | An account session in a cloud browser | Search, connect, and message; not a publishing tool | No | Free plan available |
How Do You Add Them?
The install path depends on the tool and the assistant:
- ChatGPT, one click: Of the content platforms in the table, only FeedBoss and Postbeam appeared in ChatGPT's Plugins directory when we searched.
- ChatGPT, Developer mode: Taplio, Supergrow, ContentIn, and Buffer can be connected as custom connectors instead.
- Claude: Hosted MCP servers are added as custom connectors using a server URL. Open-source projects such as
stickerdaniel/linkedin-mcp-serverrun locally on your own machine.
Is a LinkedIn MCP Server Safe?
It depends mainly on how the server reaches LinkedIn: through the front door, or through a window.
Official API Access
The tool sends you through an approved sign-in flow and uses LinkedIn API access. That is the route LinkedIn built for connected applications.
Logged-In Session Access
The tool works through your personal LinkedIn session, browser automation, cookies, or password. LinkedIn's User Agreement prohibits using "bots or other unauthorized automated methods" to create, comment on, like, share, or re-share posts. The most popular open-source LinkedIn MCP server we checked acknowledges this in its own README and warns that accounts using automated tools can be restricted or banned.
How Often Do Session-Based Tools Cause Account Restrictions?
We found no reliable public data. Rather than repeat an untraceable percentage, use two simple checks when evaluating a server:
- Does it request your LinkedIn password or session cookie? If yes, it is session-based.
- Does it redirect you to a sign-in page where you approve an app? If yes, it is using an application's official access.
How to Choose a LinkedIn MCP Server
Start with the job you actually need to do.
If You Want to Publish Your Own LinkedIn Content
Choose a content platform that works through the official API. Then compare:
- Voice support
- Quality checking
- Images
- Review flow
- Installation method
- Scheduling
If LinkedIn Is One Step in a Bigger Automation
Use an automation hub if you are comfortable building the review and scheduling flow yourself.
If You Need Prospect Data or Outreach at Volume
You are mainly looking at session-based services, browser automation, and scrapers. Know LinkedIn's terms before using those tools, and never use an account you cannot afford to lose.
Questions to Ask Any LinkedIn MCP Vendor
- How does the tool reach LinkedIn?
- Does it ever need my LinkedIn password or cookies?
- Can it publish without my approval?
- Does it post to my personal profile, a company page, or both?
- Which assistants can install it, and how?
- Can I disconnect it from one clear place?
- What does it cost per month and per action?
Where FeedBoss Fits
FeedBoss is a content platform that uses LinkedIn's official API. Its focus is the writing and publishing workflow around your own content.
- Your voice: Drafts can use your voice profile and knowledge base instead of relying only on the assistant's default writing style.
- A slop check: FeedBoss flags lines that read as AI-written and suggests a fix for each. None of the other servers in the table describe flagging AI-sounding lines.
- More than text: On-brand images, your analytics, and your posting streak, with scheduling through LinkedIn's official API.
- Easy installation: One click from ChatGPT's Plugins directory, or one server URL in Claude, Cursor, and other MCP apps.
- A preview before every schedule: Before a post is scheduled, FeedBoss shows the account, the scheduled time, and the opening of the post, then asks for your approval. You can cancel any time before the post goes out.
- A full content platform in the app: Carousels, slide decks, video, analytics, and a content calendar.
What FeedBoss is not designed for is outreach. It does not provide LinkedIn messages, prospect search, or scraping. For those jobs, use a different type of LinkedIn MCP server.
Every format in one platform, plus an MCP server for your assistant.
Start growing freeHow to Set Up a LinkedIn MCP Server
For any hosted server that uses an official sign-in, setup is usually short. For FeedBoss:
- ChatGPT: Install FeedBoss from the Plugins directory and approve the connection. The ChatGPT plugin guide has the full walkthrough.
- Claude: Add the FeedBoss MCP server URL,
https://server.feedboss.ai/api/v1/mcp, as a custom connector. The Claude setup guide includes the full steps and troubleshooting. - Claude Code, Cursor, and other MCP apps: Use the same remote server URL. The setup docs include the commands.
- Developers: The MCP server reference lists every available tool, the OAuth details, and the scheduling confirmation behavior.
Conclusion
"LinkedIn MCP server" can describe several very different kinds of tools, and LinkedIn itself does not publish one.
If your goal is creating and publishing your own LinkedIn content, start with tools that work through LinkedIn's official API. Then compare the parts that affect the final post, such as your voice, quality checks, images, approval flow, and installation. If your goal is prospecting, scraping, or outreach, you are looking at a different category of tools with different risks and tradeoffs.
For help making AI-assisted posts sound more natural, see the guides on what AI slop is and how to humanize AI writing.
Pick the server that does your job, and make sure it comes in through the front door.
FAQs
1. Can a LinkedIn MCP server post for me?
Many can. Content platforms such as FeedBoss can draft and schedule LinkedIn posts from an AI assistant, and some also require your approval before the post is scheduled or published. Scrapers and outreach servers often focus on profiles, jobs, messages, or prospect data instead of publishing your own content.
2. Is using a LinkedIn MCP server against LinkedIn's terms?
Not if the server uses an app with official API access after you sign in, which is LinkedIn's intended integration route. Servers that automate your logged-in session conflict with LinkedIn's User Agreement, which prohibits bots and unauthorized automated methods for posting.
3. Will a LinkedIn MCP server get my account banned?
An official-API server should not put your account at risk simply for posting through that approved route. Session-based servers carry more risk, and we found no reliable public data showing how often accounts get restricted.
4. Is there a free LinkedIn MCP server?
Yes. stickerdaniel/linkedin-mcp-server is free and open source but session-based. Buffer offers its MCP connection on every plan, including Free, and Composio and Zapier have free tiers with usage limits. Content platforms with voice and quality features, including FeedBoss, require a paid plan.
5. What is the difference between an MCP server and a ChatGPT plugin?
A ChatGPT plugin is an MCP-based app listed in ChatGPT's Plugins directory, which means users can install it in one click after OpenAI's review. An MCP server that is not listed in the directory can still be connected manually by URL as a custom connector. FeedBoss offers both.