LinkedIn MCP Server

A FastMCP server for LinkedIn automation that provides comprehensive LinkedIn API integration with support for creating, updating, and deleting posts with media (images, videos, documents).

🚀 Features

Core Tools

• create_post: Create LinkedIn posts with text, images, videos, or documents
• update_post: Update existing LinkedIn post text
• delete_post: Delete LinkedIn posts (supports both UGC and Share formats)

Media Support

• TEXT: Simple text posts
• IMAGE: Image uploads (JPG, PNG, GIF)
• VIDEO: Video uploads (MP4, MOV, AVI) with extended timeout support
• DOCUMENT: Document uploads (PDF, DOC, DOCX, PPT, PPTX)

Advanced Features

• Intelligent timeout handling (up to 10 minutes for large files)
• Dynamic video processing wait times based on file size
• Proper URN encoding for post operations
• Comprehensive error handling and progress feedback
• Support for both UGC posts and legacy Share posts

📋 Prerequisites

• Python 3.10 or higher
• uv package manager
• LinkedIn Developer Account with API access
• Valid LinkedIn Access Token with required permissions

Required LinkedIn API Permissions

• w_member_social: Write access to member's social posts
• (Optional) r_liteprofile: Read basic profile information

🛠 Installation

1. Clone or download the project

   mkdir linkedin-mcp
   cd linkedin-mcp
   

2. Install dependencies

   uv init
   uv add fastmcp>=2.12.4 requests
   

3. Set up LinkedIn API credentials

   • Create a LinkedIn Developer App at LinkedIn Developer Portal
   • Generate an Access Token with w_member_social permission
   • Get your Member URN (personal LinkedIn profile URN)

⚙️ Configuration

Update the following variables in main.py with your LinkedIn credentials:

# LinkedIn API Credentials (REPLACE WITH YOUR VALUES)
LINKEDIN_ACCESS_TOKEN = "YOUR_ACCESS_TOKEN_HERE"
LINKEDIN_MEMBER_URN = "urn:li:person:YOUR_MEMBER_ID"
LINKEDIN_API_BASE = "https://api.linkedin.com/v2"

How to Get Your Credentials

1. Access Token:

   • Go to LinkedIn Developer Portal
   • Create an app and request w_member_social permission
   • Generate access token from your app settings

2. Member URN:

   • Use LinkedIn API /me endpoint to get your profile
   • Extract the URN from the response
   • Format: urn:li:person:XXXXXXXXXX

🚀 Usage

Starting the Server

# Start in development mode
uv run fastmcp dev main.py

The server will start on http://localhost:6274 with MCP Inspector available for testing.

Tool Examples

1. Create Text Post

create_post(
    content="Hello LinkedIn! This is my first MCP post.",
    media_type="TEXT"
)

2. Create Image Post

create_post(
    content="Check out this amazing image!",
    media_type="IMAGE",
    file_path="C:/path/to/your/image.jpg"
)

3. Create Video Post

create_post(
    content="Watch this cool video!",
    media_type="VIDEO",
    file_path="C:/path/to/your/video.mp4"
)

4. Create Document Post

create_post(
    content="Read my latest report",
    media_type="DOCUMENT",
    file_path="C:/path/to/your/document.pdf"
)

5. Update Post

update_post(
    post_id="urn:li:ugcPost:1234567890",
    new_text="Updated post content"
)

6. Delete Post

delete_post(post_id="urn:li:ugcPost:1234567890")
# or for legacy shares
delete_post(post_id="urn:li:share:1234567890")

⏱️ Timeout Configuration

The server uses intelligent timeout handling:

• TEXT posts: 30 seconds
• IMAGE posts: 2 minutes
• VIDEO posts: 10 minutes
• DOCUMENT posts: 10 minutes

File upload timeouts are dynamic based on file size:

• Files < 10MB: 2 minutes
• Files 10-50MB: 5 minutes
• Files > 50MB: 10 minutes

🔧 Technical Details

Media Upload Process

1. Registration: Register upload with LinkedIn to get upload URL and asset URN
2. Binary Upload: Upload the actual file data to LinkedIn's storage
3. Processing Wait: Wait for LinkedIn to process the media (especially videos)
4. Post Creation: Create the LinkedIn post with the uploaded media

Error Handling

The server provides comprehensive error handling:

• File existence validation
• Media type validation
• Upload failure recovery
• Timeout management
• LinkedIn API error parsing

URN Handling

The server properly handles different LinkedIn URN formats:

• UGC Posts: urn:li:ugcPost:{id}
• Legacy Shares: urn:li:share:{id}
• Proper URL encoding for API calls

📁 Project Structure

linkedin-mcp/
├── main.py              # Main MCP server code
├── pyproject.toml       # Python project configuration
├── README.md           # This file
└── .env               # Environment variables (optional)

🔍 API Reference

create_post(content, media_type="TEXT", file_path=None)

Create a new LinkedIn post with optional media.

Parameters:

• content (str): Text content of the post
• media_type (str): Type of media - "TEXT", "IMAGE", "VIDEO", or "DOCUMENT"
• file_path (str, optional): Absolute path to media file

Returns:

• Dict with success status, post URN, and any error details

update_post(post_id, new_text)

Update the text content of an existing UGC post.

Parameters:

• post_id (str): Post URN (e.g., "urn:li:ugcPost:1234567890")
• new_text (str): New text content

Returns:

• Dict with success status and operation details

delete_post(post_id)

Delete a LinkedIn post (supports both UGC and Share formats).

Parameters:

• post_id (str): Post URN (e.g., "urn:li:ugcPost:1234567890" or "urn:li:share:1234567890")

Returns:

• Dict with success status and operation details

🚨 Important Notes

Timeout Behavior

• If you see a timeout error but the post appears on LinkedIn, the upload succeeded
• LinkedIn processes uploads in the background, so timeouts may occur before final confirmation
• Check your LinkedIn profile to verify post creation

File Size Limits

• LinkedIn has file size limits for different media types
• Large videos (>100MB) may take the full 10-minute timeout
• Consider compressing large files before upload

API Rate Limits

• LinkedIn has rate limits on API calls
• Space out your requests to avoid hitting limits
• The server includes appropriate delays for video processing

🐛 Troubleshooting

Common Issues

1. "ACCESS_DENIED" errors

   • Verify your access token has the required permissions
   • Check that your LinkedIn app has w_member_social permission

2. "File not found" errors

   • Use absolute file paths
   • Ensure file exists and is accessible

3. Upload timeouts

   • Check your internet connection
   • Try smaller files first
   • Remember: timeout doesn't always mean failure

4. Invalid URN format

   • Post IDs must start with urn:li:ugcPost: or urn:li:share:
   • Get correct URNs from post creation responses

Debug Mode

Start the server with debug logging:

uv run fastmcp dev main.py --debug

📄 License

This project is provided as-is for educational and development purposes. Please ensure compliance with LinkedIn's API Terms of Service.

🤝 Contributing

Feel free to submit issues, feature requests, or pull requests to improve this LinkedIn MCP server.

📞 Support

For questions or issues:

1. Check the troubleshooting section above
2. Review LinkedIn's API documentation
3. Test with the MCP Inspector at http://localhost:6274

Note: Replace all placeholder credentials in main.py with your actual LinkedIn API credentials before running the server.