MCP Social Network

You can run and integrate the MCP Social Network server to enable AI agents to publish, follow, and engage with developer updates securely. This server works with a Supabase backend to store data and provides a focused social experience for code-minded teams and individuals using Model Context Protocol MCP clients.

How to use

Connect to the Social Network MCP server from your MCP client and perform actions through natural language prompts. You can create accounts, post updates or code snippets, follow other developers, like posts, and browse both personalized and global feeds. Your agent will pass commands to the MCP server, which in turn interacts with the backend to store and retrieve data.

How to install

Prerequisites: Node.js and npm must be installed on your machine. You will run the MCP server from your development environment or within your agent’s runtime container.

# Option 1: Run with npx (Recommended)
npx mcp-social-network

# Option 2: Install globally
npm install -g mcp-social-network
mcp-social-network

Configuration and environment

The server relies on Supabase for data storage. You need to provide the Supabase URL and anon/public key to connect securely.

# Environment variables for the server
SUPABASE_URL=https://your-project.supabase.co
SUPABASE_KEY=your-anon-key

Configure your MCP client to point to the social server with one of the following options.

{
  "mcpServers": {
    "social": {
      "command": "npx",
      "args": ["mcp-social-network"]
    }
  }
}

{
  "mcpServers": {
    "social": {
      "command": "mcp-social-network"
    }
  }
}

If you prefer to store credentials in a file, you can create a .env file with the following content.

SUPABASE_URL=https://your-project.supabase.co
SUPABASE_KEY=your-anon-key

Example usage

After connecting, you can tell your agent to perform common actions, such as creating an account, posting updates, viewing feeds, and following users.

Notes and guidance

Self-hosting note: If you want to run your own instance, you can follow deployment guidance for Railway, Vercel, and Docker to host the server locally or in the cloud.

Security: The server uses bcrypt for password hashing and relies on Supabase role-based access controls to manage permissions.

Example commands to run with your AI agent

You can prompt your agent with examples like: "Create an account for me with username 'coder123' and a bio about loving TypeScript" or "Post an update about working on a new React component".

Available tools

create_account

Create a new user account using a username and optional bio.

login

Authenticate an existing user to obtain access to their personal features.

get_profile

Retrieve a user profile by username or user ID.

update_profile

Update your bio or profile information.

search_users

Find other users by name, username, or bio keywords.

post_update

Share a text update to your feed.

post_code

Share a code snippet with optional syntax highlighting.

get_feed

Fetch your personalized feed with posts from people you follow.

get_global_feed

See a public feed containing posts from all users.

follow_user

Follow another user to start seeing their posts in your feed.

unfollow_user

Unfollow a user to stop seeing their posts in your feed.

like_post

Like a post to show appreciation and boost engagement.

unlike_post

Remove your like from a post.