Is there a Claude connector for LinkedIn?

Yes. In Claude, open Customize, then Connectors, add a custom connector named ReachHunt with https://mcp.reachhunt.ai/mcp, and sign in. ChatGPT works the same way through Apps in developer mode, so you get a ChatGPT connector for LinkedIn as well.

Do I need to code anything?

No. You paste one URL and sign in. Developers can also use the same tools from Claude Code, Codex CLI or Cursor.

Can Claude send LinkedIn connection requests and messages?

Yes, through campaigns and the inbox. Every action that reaches a real person is marked, so Claude shows it to you and waits for your OK before it runs.

Is it safe for my LinkedIn account?

Yes. Every action follows the same warm-up, smart limits and human-like pacing as the ReachHunt app, sending tools ask before they run, and your AI app can’t delete anything.

Your AI app never sees a password or an API key. Sign-in runs on OAuth, tokens are stored as hashes, and the workspace key lives only inside the encrypted grant.

Does it work with X (Twitter) too?

Yes. The same server covers X: DMs, lead lists from followers and retweets, and DM campaigns. One connection gives your AI app both networks.

How much does it cost?

The MCP server comes with every ReachHunt plan at no extra cost, and you can try it free for 14 days with no card required.

Remote MCP server

LinkedIn MCP server for Claude, ChatGPT and Cursor.

Sign in once and your AI app works in LinkedIn through the accounts you already use. It searches people, builds lead lists, sends connection requests and messages, runs campaigns and answers your inbox. No code, nothing to host.

Start free Setup guide

Server URLhttps://mcp.reachhunt.ai/mcp

Works in Claude
No code
X included

Claude

ReachHunt connected

Find heads of sales at SaaS companies in Berlin and start a connect-then-message campaign.

linkedin_search_optionsLocation Berlin, industry Software

lead_lists_createList “Berlin sales leaders” created

lead_lists_start_import312 people imported from the search

campaigns_createConnection request, then a message after 2 days

Done. 312 heads of sales are in the list. The campaign sends a connection request from Jane Cooper, then a short message once they accept.

Claude wants to use campaigns_start

Sends connection requests to 312 people from Jane Cooper

AllowDeny

The basics

What a LinkedIn MCP server does for your AI app.

MCP (Model Context Protocol) is the open standard AI apps use to call outside tools. Claude, ChatGPT, Cursor, Claude Code, Codex and Gemini CLI all support it. A LinkedIn MCP server gives your AI app tools for LinkedIn, so instead of clicking through searches and inboxes, you ask in plain words and the app does the work.

LinkedIn’s own API doesn’t let apps search people or send messages for outreach. ReachHunt fills that gap: the server runs in the cloud at mcp.reachhunt.ai, you sign in with your ReachHunt account, and every action goes through the LinkedIn accounts connected to your workspace, with smart limits keeping them healthy.

The same server covers X (Twitter), so one connection gives Claude or ChatGPT both networks, one inbox and one pipeline.

What it can do

Your whole LinkedIn outreach, from one chat.

Ask for a result, not a series of clicks. Anything that reaches real people waits for your OK first.

Find the right people
Search LinkedIn by keyword, title, company, location and industry, or paste a search link, then import your connections, followers, post likers, commenters or group members.
Connect and message
Run campaigns that send a connection request, then a message once they accept, or a free InMail to Open Profile members.
Answer your inbox
Read conversations and invitations, draft replies, send messages and schedule follow-ups for later.
Run campaigns
Create a sequence, start, pause or stop it, and see who replied and who is next.
Put an AI agent on it
Train an agent on your company, test it, and let it answer LinkedIn replies around the clock.
Keep the pipeline current
Tag conversations, set statuses, notes, owners and deal values, and read your pipeline.

Examples

Ask in plain words. The tools do the rest.

Real requests your AI app can handle in ReachHunt, in plain words.

“Find heads of sales at SaaS companies in Berlin and add them to a new list.”
Searches LinkedInBuilds a list
“Import everyone who commented on my last LinkedIn post.”
Imports commentersBuilds a list
“Send a connection request to this list, then a short message two days after they accept.”
Writes the sequenceStarts the campaign
“Summarize my unread LinkedIn messages and draft replies to the hot leads.”
Reads your inboxDrafts replies
“Accept the invitations from founders and say thanks to each one.”
Reads invitationsSends replies
“Which campaign got the most replies this week, and who replied?”
Reads campaignsLists replies
“Tag everyone who asked for a demo as Hot and assign them to Sam.”
Finds the chatsAdds a tagAssigns a teammate
“Train an agent on our FAQ and put it on my LinkedIn account.”
Trains an agentPuts it live

Setup

Connect in two minutes. No keys to copy.

Paste the server URL, sign in to ReachHunt in the browser window that opens, and pick your workspace. You need full access to Developer in the workspace and an active plan.

Claude
On claude.ai or the desktop app, open Customize, then Connectors. Select + Add, then Add custom connector, name it ReachHunt and paste the URL. Select Connect, pick your workspace, then turn ReachHunt on from the + menu in a chat.
Server URL
```
https://mcp.reachhunt.ai/mcp
```
ChatGPT
Open Settings, then Apps, then Advanced settings, and turn on Developer mode. Back in Apps, select Create app, name it ReachHunt, paste the URL as the MCP server URL, choose OAuth and connect your workspace.
MCP server URL
```
https://mcp.reachhunt.ai/mcp
```
Cursor
Add ReachHunt to ~/.cursor/mcp.json, or .cursor/mcp.json inside one project. Then open Cursor Settings, then MCP, and select the sign-in button next to reachhunt.
mcp.json
```
{
  "mcpServers": {
    "reachhunt": {
      "url": "https://mcp.reachhunt.ai/mcp"
    }
  }
}
```
Claude Desktop config
Apps that only run local servers can reach ReachHunt through mcp-remote. Add this to claude_desktop_config.json. It needs Node.js.
claude_desktop_config.json
```
{
  "mcpServers": {
    "reachhunt": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://mcp.reachhunt.ai/mcp"]
    }
  }
}
```
Claude Code
Add the server from your terminal. Start Claude Code, type /mcp, choose reachhunt and select Authenticate.
Terminal
```
claude mcp add --transport http reachhunt https://mcp.reachhunt.ai/mcp
```
Codex CLI
Add the server, then log in. Gemini CLI, VS Code with GitHub Copilot and Windsurf are covered in the docs.
Terminal
```
codex mcp add reachhunt --url https://mcp.reachhunt.ai/mcp
codex mcp login reachhunt
```

Safety

Your AI app asks before anything goes out.

An AI app gets real power in your inbox, so the guardrails sit on the server, not in a prompt.

Sending asks first
Connection requests, messages and campaign starts are marked as actions that reach real people, so Claude and ChatGPT wait for your OK.
Nothing can be deleted
Your AI app can’t delete conversations, lead lists, campaigns, agents, templates, tags or statuses.
Account safety built in
Every LinkedIn account keeps its warm-up, smart limits and human-like pacing, whatever sends the message.
Sign-in, not keys
You connect with OAuth and pick a workspace. No password or API key ever reaches the AI app.

Compare

ReachHunt vs a typical open-source LinkedIn MCP server.

ReachHunt MCP compared with a typical self-hosted LinkedIn MCP server
	ReachHunt MCP	Self-hosted LinkedIn MCP
Setup	Paste a URL and sign in	Clone, configure and run it
Hosting	Cloud, nothing to run	Your machine or server
Search people and build lead lists		Some
Connection requests and messages		Some
Campaigns with follow-ups
AI agents that answer replies
Shared team inbox and pipeline
Warm-up and smart limits
X (Twitter) in the same server

FAQ

LinkedIn MCP questions, answered.

Still have a question?

Book a demo

A remote MCP server at https://mcp.reachhunt.ai/mcp that lets AI apps work in your ReachHunt workspace. On LinkedIn, your AI app can search people, build lead lists, run campaigns with connection requests and messages, answer your inbox and manage AI agents.

Keep reading

Now live

Your next customer is one DM away.

Connect an account, build a list and switch on your AI agent — in about the time it took to read this page.

Start free Book a demo

14-day free trial
No card required
2-minute setup

LinkedIn MCP server for Claude, ChatGPT and Cursor.

What a LinkedIn MCP server does for your AI app.

Your whole LinkedIn outreach, from one chat.

Find the right people

Connect and message

Answer your inbox

Run campaigns

Put an AI agent on it

Keep the pipeline current

Ask in plain words. The tools do the rest.

Connect in two minutes. No keys to copy.

Claude

ChatGPT

Cursor

Claude Desktop config

Claude Code

Codex CLI

Your AI app asks before anything goes out.

Sending asks first

Nothing can be deleted

Account safety built in

Sign-in, not keys