LI
OfficialDeveloper Tools

LinkedIn MCP Server: Write, Edit, Schedule Posts

Create, edit, and schedule LinkedIn posts from ChatGPT and Claude using LiGo's MCP server for streamlined publishing.

View on GitHub
Quick Install
npx -y @ertiqah/linkedin-mcp-runner

Overview

This MCP (Model Context Protocol) server connects ChatGPT and Claude assistants to your public LinkedIn context so they can create, edit, and schedule LinkedIn posts on your behalf. It acts as a bridge between a GPT-based UI and LiGo’s LinkedIn tooling: the assistant can fetch your recent public posts, analyze engagement and tone, generate rewrites in your voice, and submit or schedule new posts via the LiGo platform.

For developers, the server exposes the MCP endpoints that LLM frontends use to request contextual data and perform actions. That makes it simple to plug an assistant into a live workflow: ask a Claude or ChatGPT-powered co-pilot to “analyze my last 5 posts” or “rewrite this draft to match my tone,” and the server supplies the LinkedIn context and posting capabilities to support the request.

Features

  • Pull public LinkedIn profile data (posts, timestamps, basic engagement metrics)
  • Analyze recent posts for tone, structure, and engagement patterns
  • Generate and rewrite post drafts that match your voice
  • Create and publish LinkedIn posts programmatically
  • Schedule posts for future publication
  • Works with Claude (desktop integration) and ChatGPT (CustomGPT)
  • Public leaderboard to surface recent MCP-created posts (optional, public feed)
  • Integrates with the broader LiGoSocial ecosystem (analytics, commenting, CRM)

Installation / Configuration

Prerequisites: git, Node.js (or the runtime specified in the repo), and a LiGo account to authorize the integration.

Quick local install (example):

# clone the repository
git clone https://github.com/ertiqah/linkedin-mcp-runner.git
cd linkedin-mcp-runner

# install dependencies (example; repo may use npm or yarn)
npm install

# copy environment template and edit
cp .env.example .env
# Edit .env to include required values:
# - LIGO_API_KEY (or CLIENT_ID/CLIENT_SECRET)
# - MCP_HOST (optional)
# - CALLBACK_URL or similar auth redirect

Start the server:

# development
npm run dev

# or production
npm start

If you are deploying to a hosting service, make sure your environment variables are set and HTTPS is enabled (MCP integrations expect secure callbacks).

Claude integration (high level)

  1. In LiGo dashboard, navigate to Integrations → Claude and click “Generate Installation Command.”
  2. Copy the terminal command and run it on the machine where your Claude app can access the MCP server (this registers the server with Claude’s desktop app).
  3. Open Claude and begin interacting with your assistant; prompts can ask the assistant to fetch or analyze your LinkedIn data.

ChatGPT / CustomGPT integration (high level)

  1. In LiGo dashboard, go to Integrations → ChatGPT (CustomGPT).
  2. Authenticate LiGo when prompted and follow the CustomGPT setup flow.
  3. Use the CustomGPT to issue prompts directly — no local install needed.

Available Resources

The MCP server exposes a small set of endpoints the assistant uses to read context and perform actions. The exact routes may vary by release; below is a representative reference to get started.

EndpointMethodPurpose
/mcp/profileGETReturn public profile metadata and basic stats
/mcp/posts?limit=NGETFetch recent N public posts with engagement metrics
/mcp/postsPOSTCreate a new post (payload includes content, visibility, optional schedule_at)
/mcp/posts/{id}PATCHEdit an existing draft or scheduled post
/mcp/posts/{id}/schedulePOSTSchedule a draft to publish at a future timestamp
/mcp/posts/{id}/publishPOSTPublish immediately

Example create post request (curl):

curl -X POST "https://your-mcp-host/mcp/posts" \
  -H "Authorization: Bearer $LIGO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "Draft post text created by the assistant.",
    "visibility": "public",
    "schedule_at": "2026-04-15T14:00:00Z"
  }'

Example response (abridged):

{
  "id": "post_12345",
  "status": "scheduled",
  "schedule_at": "2026-04-15T14:00:00Z",
  "url": "https://linkedin.com/posts/..."
}

Use Cases

  • Analyze recent performance

    • Prompt: “Analyze my last 5 posts. Which got the most traction and why?”
    • Result: Assistant pulls the last 5 posts, computes engagement metrics, and summarizes patterns.

  • Rewrite a draft in your voice

    • Prompt: “Rewrite this draft to match the tone of my recent posts and make the hook stronger.”
    • Result: System fetches sample posts to infer voice then provides rewritten options.

  • Create and schedule a post

    • Prompt: “Draft a post about product updates and schedule it for Tuesday at 10am PT.”
    • Result: Assistant creates the draft, returns the preview, and schedules via the MCP endpoint.

  • Quick publish from chat

    • Prompt: “Polish this 3-line post and publish now.”
    • Result:
Tags:developer-tools