Taplio LinkedIn MCP Server

Remote server

The Taplio LinkedIn MCP Server is a Model Context Protocol (MCP) server that connects AI tools directly to your Taplio account and, through it, to your LinkedIn presence. It gives AI agents, assistants, and chat clients the ability to draft, schedule, and publish LinkedIn posts, browse what you have already posted, research inspiration from other creators, and read your analytics, all through natural language.

Point your MCP host at the Taplio server and you can ask things like:

• "Draft a LinkedIn post about our new MCP launch and schedule it for Tuesday 9am."
• "Show me my last 10 posts and which one got the most impressions."
• "Find 20 high-performing posts about linkedin ghostwriting from creators under 50k followers."
• "How are my followers and engagement trending over the last 30 days?"

Use cases

• Turning raw ideas into publish-ready LinkedIn posts, on your own voice and topics.
• Automating draft creation, scheduling, and publishing from any MCP-capable client.
• Building AI-assisted content workflows and reporting on top of your real LinkedIn data.
• Extracting insights and analytics from your posting history without leaving your assistant.

Contents

• Remote Taplio MCP Server
• Prerequisites
• Installation
  • Claude Code
  • Claude Desktop / claude.ai connectors
  • VS Code
  • Cursor
  • Windsurf
  • Generic MCP config
• Authentication
• Toolsets
• Tools
  • Profile
  • Drafts
  • Publishing
  • Posts
  • Analytics
  • Inspiration
• Recommended workflow
• Library and lifecycle model
• Limits and safety
• Troubleshooting
• Contributing
• License

Remote Taplio MCP Server

Taplio is offered as a hosted, remote MCP server. There is nothing to install or run locally: your MCP host connects to the Taplio endpoint over HTTP and authenticates with OAuth.

Because the server is remote and OAuth-based, you never copy a secret into a config file. Access is tied to the Taplio account you approve in the browser, and can be revoked from Taplio at any time.

Prerequisites

1. An active Taplio account with LinkedIn connected.
2. An MCP-capable host (Claude Code, Claude Desktop, VS Code with Copilot, Cursor, Windsurf, or any client that speaks MCP over HTTP).
3. A browser available on first connect, to complete the OAuth authorization.

Installation

Claude Code

claude mcp add --transport http taplio https://mcp.taplio.com

Then run /mcp inside Claude Code and authorize taplio in the browser when prompted.

Claude Desktop / claude.ai connectors

In Claude Desktop or on claude.ai, open Settings -> Connectors -> Add custom connector, then enter:

• Name: Taplio
• URL: https://mcp.taplio.com

Approve the OAuth screen. The Taplio tools then appear in the tool picker.

VS Code

Add to your user or workspace MCP configuration (.vscode/mcp.json):

{
  "servers": {
    "taplio": {
      "type": "http",
      "url": "https://mcp.taplio.com"
    }
  }
}

Cursor

In ~/.cursor/mcp.json (or Settings -> MCP -> Add):

{
  "mcpServers": {
    "taplio": {
      "url": "https://mcp.taplio.com"
    }
  }
}

Windsurf

In ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "taplio": {
      "serverUrl": "https://mcp.taplio.com"
    }
  }
}

Generic MCP config

Any client that supports remote MCP servers over HTTP can use:

{
  "mcpServers": {
    "taplio": {
      "type": "http",
      "url": "https://mcp.taplio.com"
    }
  }
}

Authentication

The Taplio MCP Server uses OAuth 2.0. On first use your client opens a browser window where you log in to Taplio and approve access. The server then acts on behalf of that single account. No API keys, tokens, or LinkedIn credentials are stored in your client config. Revoke access at any time from your Taplio account settings.

Toolsets

The tools are grouped into functional toolsets. All of them operate on the authenticated user's own account.

Tools

Profile

get_me

Read this first to orient. Returns the user's identity and content preferences so you can tailor output to them and recognize their own posts: display name and LinkedIn @handle (a user may go by several aliases, so check both), plus ai_settings (industry, role, language, target audience, topics, keywords, description) and today's LinkedIn stats (followers, connections, profile views).

No parameters.

Drafts

create_draft

Create a new LinkedIn post draft (unpublished). Publish it later with schedule_draft or publish_draft.

get_draft

Fetch a single LinkedIn post draft by id.

list_drafts

List the user's unpublished LinkedIn post drafts.

update_draft

Update the content of an existing draft. Only drafts can be edited: published or scheduled posts cannot.

delete_draft

Permanently delete a LinkedIn post draft.

Publishing

publish_draft

Publish a draft to LinkedIn immediately. This goes live now and cannot be undone.

schedule_draft

Schedule a draft to be published automatically at a future time.

unschedule

Unschedule a scheduled post, returning it to drafts so you can edit, reschedule, or delete it.

Posts

list_posts

List the user's LinkedIn posts (status scheduled, sending, or sent), most recent first. A draft becomes a post once it is scheduled or published.

get_post

Fetch a single LinkedIn post by id. A post has status scheduled, sending, or sent; drafts are separate (use get_draft).

Analytics

get_analytics_overview

Get the user's LinkedIn analytics overview (followers, impressions, engagement). Defaults to today when no date range is given.

get_post_analytics

List the user's per-post analytics (impressions, likes, comments, shares), defaulting to the last 7 days.

Inspiration

search_inspiration

Search LinkedIn posts from other creators to use as inspiration.

Recommended workflow

1. Call get_me once at the start of a session to load the user's voice, topics, language, and stats.
2. Use search_inspiration to ground new content in what is currently working in the niche.
3. Draft with create_draft, iterate with update_draft.
4. Ship with schedule_draft (preferred) or publish_draft (immediate).
5. Report with get_analytics_overview and get_post_analytics, and cross-reference with list_posts.

Library and lifecycle model

A draft and a post are the same item at different stages of its life:

create_draft ─┐
              ├─> [ DRAFT ] ──schedule_draft──> [ scheduled ] ──(auto)──> [ sending ] ──> [ sent ]
update_draft ─┘        │                              │
                       └──────publish_draft───────────┴──(immediate)──> [ sending ] ──> [ sent ]

                       [ scheduled ] ──unschedule──> [ DRAFT ]

• Drafts live in the drafts list (list_drafts) and are the only editable stage.
• Scheduling or publishing a draft turns it into a post and moves it to the posts list (list_posts).
• unschedule moves a scheduled post back to drafts; a sent post cannot be un-sent.

Limits and safety

• Post length: 1 to 3000 characters (create_draft, update_draft).
• Scheduling: scheduled_for must be at least 2 minutes in the future, ISO 8601, UTC recommended.
• Analytics range: capped at 90 days per query; paginate with cursor for large result sets.
• Irreversible actions: publish_draft goes live immediately and delete_draft is permanent. Confirm intent before calling either.
• Scope: every tool acts only on the authenticated user's own Taplio and LinkedIn account.

Troubleshooting

• Tools do not appear: confirm the server URL is https://mcp.taplio.com and that you completed the OAuth browser step. Re-run /mcp (Claude Code) or reconnect the connector.
• 401 / auth errors: your session may have expired or been revoked. Reconnect and re-authorize in the browser.
• Cannot edit a post: only drafts are editable. If it is already scheduled, call unschedule first; if it is sent, it can no longer be changed.
• Empty analytics: newly published posts take time to accumulate metrics; also check your date range against the 90-day cap.

Contributing

This repository documents the Taplio LinkedIn MCP Server. See CONTRIBUTING.md for how to report issues with the documentation or suggest improvements. For product support, contact Taplio.

License

This project is licensed under the terms of the MIT License. See LICENSE.