How do I use mcp-server-linkedin?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-server-linkedin Publish a LinkedIn post about our new AI feature with an image" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

mcp-server-linkedin

by Huzaifa-ali

Overview Schema Source Related Servers Discussions Admin

Python

Remote

Use Cases

Content Publishing: Draft and publish LinkedIn posts from your AI assistant without switching context.
Media Sharing: Upload images and videos alongside your posts in a single command.
Link Previews: Share articles with auto-generated link preview cards.
Multi-Format Workflow: Compose text-only updates, visual content, or article shares through one unified interface.
Account Management: Authenticate, check your profile, and manage your session without leaving your editor.

Built for developers and content creators who want their AI tools to publish directly to LinkedIn.

Quick Start

Prerequisites

Python 3.10+
LinkedIn App — Follow the step-by-step setup guide or the quick version below:
- Create an app at LinkedIn Developer Portal
- Products: "Share on LinkedIn" + "Sign In with LinkedIn using OpenID Connect"
- OAuth 2.0 scopes: openid, profile, email, w_member_social
- Redirect URL: http://localhost:3000/callback

Install

uvx mcp-server-linkedin

Or with pip:

pip install mcp-server-linkedin

Related MCP server: LinkedIn MCP Server

Configuration

Set these environment variables:

LINKEDIN_CLIENT_ID=your_client_id        # From LinkedIn Developer Portal
LINKEDIN_CLIENT_SECRET=your_secret_here  # From LinkedIn Developer Portal
LINKEDIN_REDIRECT_URI=http://localhost:3000/callback  # Optional, this is the default

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "linkedin": {
      "command": "uvx",
      "args": ["mcp-server-linkedin"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id",
        "LINKEDIN_CLIENT_SECRET": "your_secret_here"
      }
    }
  }
}

{
  "mcpServers": {
    "linkedin": {
      "command": "uvx",
      "args": ["mcp-server-linkedin"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id",
        "LINKEDIN_CLIENT_SECRET": "your_secret_here"
      }
    }
  }
}

VS Code + Copilot

Add to .vscode/mcp.json in your workspace:

{
  "servers": {
    "linkedin": {
      "command": "uvx",
      "args": ["mcp-server-linkedin"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id",
        "LINKEDIN_CLIENT_SECRET": "your_secret_here"
      }
    }
  }
}

Cursor

Add to .cursor/mcp.json in your project root:

{
  "mcpServers": {
    "linkedin": {
      "command": "uvx",
      "args": ["mcp-server-linkedin"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id",
        "LINKEDIN_CLIENT_SECRET": "your_secret_here"
      }
    }
  }
}

Kiro

Add to .kiro/settings/mcp.json:

{
  "mcpServers": {
    "linkedin": {
      "command": "uvx",
      "args": ["mcp-server-linkedin"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id",
        "LINKEDIN_CLIENT_SECRET": "your_secret_here"
      }
    }
  }
}

Claude Code

claude mcp add linkedin -- uvx mcp-server-linkedin

Then set LINKEDIN_CLIENT_ID and LINKEDIN_CLIENT_SECRET in your environment.

Windsurf

Add to your Windsurf MCP configuration:

{
  "mcpServers": {
    "linkedin": {
      "command": "uvx",
      "args": ["mcp-server-linkedin"],
      "env": {
        "LINKEDIN_CLIENT_ID": "your_client_id",
        "LINKEDIN_CLIENT_SECRET": "your_secret_here"
      }
    }
  }
}

Running from Source

Replace "command": "uvx", "args": ["mcp-server-linkedin"] with:

{
  "command": "uv",
  "args": ["--directory", "/path/to/mcp-server-linkedin", "run", "mcp-server-linkedin"]
}

Authentication

Start your MCP client (Claude Desktop, Kiro, etc.)
Ask: "Authenticate with LinkedIn"
Browser opens → authorize the app → callback captured automatically
Token saved to ~/.mcp-server-linkedin/token.json
Token lasts 2 months. Re-run linkedin_auth when it expires.

Tools

Tool	Description
`linkedin_auth`	OAuth 2.0 browser-based authentication
`linkedin_get_profile`	Get your name, email, and person URN
`linkedin_logout`	Remove stored token
`linkedin_post_text`	Publish a text-only post
`linkedin_post_image`	Publish a post with an image
`linkedin_post_video`	Publish a post with a video (up to 200 MB)
`linkedin_post_article`	Publish a post with a link preview
`linkedin_delete_post`	Delete a post by ID
`linkedin_get_post_stats`	Get post analytics (requires Community Management API)
`linkedin_get_all_stats`	Get aggregated analytics (requires Community Management API)

Tool Details

Parameter	Type	Default	Description
`text`	string	required	Post content (up to ~3000 chars)
`visibility`	string	`"PUBLIC"`	`"PUBLIC"` or `"CONNECTIONS"`

Parameter	Type	Default	Description
`text`	string	required	Post caption
`image_path`	string	required	Absolute path to image (JPEG, PNG, GIF)
`visibility`	string	`"PUBLIC"`	`"PUBLIC"` or `"CONNECTIONS"`

Parameter	Type	Default	Description
`text`	string	required	Post caption
`video_path`	string	required	Absolute path to video (MP4, max 200 MB)
`visibility`	string	`"PUBLIC"`	`"PUBLIC"` or `"CONNECTIONS"`

Parameter	Type	Default	Description
`text`	string	required	Post commentary
`url`	string	required	Article URL (generates link preview)
`title`	string	`""`	Optional link preview title
`description`	string	`""`	Optional link preview description
`visibility`	string	`"PUBLIC"`	`"PUBLIC"` or `"CONNECTIONS"`

Parameter	Type	Default	Description
`post_id`	string	required	Post URN (e.g., `urn:li:ugcPost:123456`)

Architecture

src/mcp_server_linkedin/
├── config.py            # Constants, API URLs, settings dataclass
├── exceptions.py        # Typed exception hierarchy
├── server.py            # FastMCP entrypoint + tool registration
├── models/              # Frozen dataclass API response models
├── services/            # Async API client with connection pooling
├── tools/               # MCP tool implementations (validate → delegate → format)
│   ├── auth.py          # OAuth flow, profile, logout
│   ├── posting.py       # Text, image, video, article, delete
│   └── analytics.py     # Stubbed (pending API access)
└── utils/               # Token persistence

Rate Limits

Limit	Value
API requests per member per day	150
Token duration	2 months
Max video file size	200 MB

The server relays LinkedIn's rate limit errors clearly but does not enforce limits internally.

Security

Tokens stored locally at ~/.mcp-server-linkedin/token.json
No credentials in code — all secrets via environment variables
Official API only — uses w_member_social scope
Local OAuth callback — authorization code never leaves your machine
No data collection — this server sends nothing except LinkedIn API calls

Development

git clone https://github.com/Huzaifa-ali/mcp-server-linkedin.git
cd mcp-server-linkedin
uv sync --all-extras
pre-commit install

uv run mcp-server-linkedin                    # Run the server
uv run ruff check src/                        # Lint
uv run ruff format src/                       # Format
uv run mypy src/                              # Type check
uv run pytest                                 # Test

Test with MCP Inspector:

npx @modelcontextprotocol/inspector uv run mcp-server-linkedin

Contributing

See CONTRIBUTING.md for guidelines. In short:

Fork the repo and create a feature branch
Make your changes with type hints and docstrings
Run pre-commit run --all-files (must pass)
Open a pull request

License

MIT — see LICENSE.

Available Tools

10 tools

linkedin_authA

Authenticate with LinkedIn using OAuth 2.0. Opens a browser window for the user to log in and authorize. Must be called BEFORE any other LinkedIn tool if the user is not yet authenticated. Only needs to be run once — the token is saved and lasts 2 months.

Inputs: no named parameters. Returns: result

linkedin_delete_postA

Delete an existing LinkedIn post. Use when the user wants to remove a post they previously published. Requires: the post URN/ID (e.g., urn:li:ugcPost:123456 or urn:li:share:123456). This action is irreversible.

Inputs: post_id (required). Returns: result

linkedin_get_all_statsA

Get aggregated analytics across ALL LinkedIn posts (total impressions, clicks, likes, comments, shares). NOTE: Currently requires Community Management API access (r_member_postAnalytics scope) which must be applied for separately. Returns instructions on how to gain access if not available.

Inputs: metric, end_date, start_date, aggregation. Returns: result

linkedin_get_post_statsB

Get analytics (impressions, clicks, likes, comments, shares) for a specific LinkedIn post. NOTE: Currently requires Community Management API access (r_member_postAnalytics scope) which must be applied for separately at developer.linkedin.com. Returns instructions on how to gain access if not available.

Inputs: post_id (required), metric, end_date, start_date, +1 more. Returns: result

linkedin_get_profileA

Get the currently authenticated LinkedIn user's profile: name, email, person URN, and profile picture URL. Use this to confirm who is logged in or to retrieve the user's identity. Requires prior authentication.

Inputs: no named parameters. Returns: result

linkedin_logoutA

Log out of LinkedIn by deleting the stored OAuth token. Use when the user wants to switch accounts or revoke access. After logout, linkedin_auth must be called again before using any other tools.

Inputs: no named parameters. Returns: result

linkedin_post_articleA

Publish a LinkedIn post with a link/article preview card. Use when the user wants to share a URL with an auto-generated preview (title, description, thumbnail). Best for sharing blog posts, news articles, or web pages. Requires: text commentary AND a URL. Optional: custom title, description, visibility (PUBLIC or CONNECTIONS).

Inputs: url (required), text (required), title, visibility, +1 more. Returns: result

linkedin_post_imageA

Publish a LinkedIn post with an attached image. Use when the user wants to share a photo, infographic, screenshot, or any image file alongside text. Requires: text caption AND absolute file path to an image (JPEG, PNG, or GIF). Optional: visibility (PUBLIC or CONNECTIONS).

Inputs: text (required), image_path (required), visibility. Returns: result

linkedin_post_textA

Publish a text-only post to LinkedIn. Use when the user wants to share a written update, thought, or announcement WITHOUT any image, video, or link preview. Requires: text content. Optional: visibility (PUBLIC or CONNECTIONS).

Inputs: text (required), visibility. Returns: result

linkedin_post_videoA

Publish a LinkedIn post with an attached video. Use when the user wants to share a video file (MP4, max 200 MB) alongside text. Requires: text caption AND absolute file path to a video file. Optional: visibility (PUBLIC or CONNECTIONS).

Inputs: text (required), video_path (required), visibility. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A4.1/5.0

Scored 2026-07-28 10:10 across 10 tools

Disambiguation5/5

Each tool targets a distinct action: authentication, profile retrieval, logout, various post types (text, image, video, article), deletion, and post stats. No overlapping functionality.

Naming Consistency5/5

All tools follow a consistent 'linkedin_verb_noun' pattern in snake_case (e.g., linkedin_post_text, linkedin_get_profile). Minor exception: linkedin_auth omits a noun but is still clear.

Tool Count5/5

With 10 tools, the set covers authentication, profile, posting (four types), deletion, and two levels of stats. Each tool serves a clear purpose, and no tool seems redundant or out of place.

Completeness3/5

Covers core posting and basic profile/stats, but notable gaps exist: no ability to list or edit posts, no comment operations, and stats tools require special API access not guaranteed. For a content-focused toolset, updates and retrieval would be expected.

Resources

Related MCP Connectors

PostPenA
ai.postpen
Schedule and publish LinkedIn posts to profiles and company pages from AI agents.
App Automation Marketing Social Media
161
https://mcp.postpen.ai
LinkedIn by 1Social
dev.1social
Publish and schedule LinkedIn posts from your AI assistant.
Social Media App Automation
1
https://mcp.1social.dev/mcp/linkedin
mcp
com.curviate
LinkedIn actions for AI agents: search, messaging, posts and invites, as hosted MCP tools.
Social Media Communication
1
https://app.curviate.com/mcp
Reach MCP — LinkedIn for AI agentsA
com.reachmcp
LinkedIn for AI agents: inbox, invitations, Sales Navigator search, posts, jobs, quotas, webhooks.
Social Media CRM
661
https://app.reachmcp.com/mcp

Related MCP Servers

LinkedIn MCP Server
Social Media Communication
ahmedekramalsada
A
license
Not graded
quality
C
maintenance
Enables AI agents to interact with LinkedIn for posting, commenting, liking, and managing connections via OAuth2 authentication.
Updated 24 days ago (2026-10-22 22:44 UTC)
18 npm
MIT
LinkedIn MCP Server
Social Media Communication
quinnjr
A
license
B
quality
C
maintenance
Enables AI agents to manage LinkedIn profiles, posts, connections, skills, education, and certifications through the LinkedIn API.
Updated 3 months ago (2026-08-03 15:06 UTC)
18
176 npm
64
MIT
LinkedIn MCP Server
Social Media Marketing
himanshu31shr
A
license
B
quality
C
maintenance
Enables AI agents with read/write access to LinkedIn API, including profile, posts, media, organizations, comments, reactions, and analytics.
Updated 5 months ago (2026-06-21 19:15 UTC)
20
7 npm
MIT
linkedin-mcp
Social Media Autonomous Agents
2060-io
A
license
A
quality
C
maintenance
Enables AI agents to publish posts, images, comments, and reactions to LinkedIn as the authenticated user, with built-in safety features like daily budgets and deduplication.
Updated 5 months ago (2026-06-22 20:57 UTC)
9
18 npm
1
Apache 2.0

Use Cases

Quick Start

Prerequisites

Install

Configuration

Claude Desktop

VS Code + Copilot

Cursor

Kiro

Claude Code

Windsurf

Running from Source

Authentication

Tools

Tool Details

Architecture

Rate Limits

Security

Development

Contributing

License

Available Tools

Tool Schema Changelog

TDQS

Resources

Related MCP Connectors

Related MCP Servers

LinkedIn MCP Server

LinkedIn MCP Server

LinkedIn MCP Server

linkedin-mcp

Appeared in Searches