mcp-link

MCP Link Server is a Python-based SSE server that provides a local tool execution environment for MCP-aware AI agents. It exposes a secure, configurable set of tools (file operations, system commands, browser automation, Docker sandboxing, and more) that agents can discover and execute under user-approved permissions. Clients connect via the SSE endpoint, and you can control which tools are available and how calls are authorized. To get started, clone the repository, install Python dependencies, and launch the server. Once running, connect your MCP client or browser extension to the server URL to begin tool discovery and task execution. The extension and the MCP clients will browse the available tools, request execution, and handle results through the SSE stream, all under your local control.

Prerequisites:

• Python 3.9+ installed on your system
• Git installed
• Access to the mcp-link-server repository (clone from GitHub)

Installation steps:

1. Clone the repository git clone https://github.com/AuraFriday/mcp-link-server.git cd mcp-link-server

2. (Optional) Create a virtual environment python -m venv venv source venv/bin/activate # macOS/Linux venv\Scripts\activate # Windows

3. Install dependencies pip install -r requirements.txt

4. Run the server python friday.py

5. Verify access

   • The server should be available at the configured SSE endpoint, typically https://127-0-0-1.local.aurafriday.com:31173/sse if using the default local setup.

Notes:

• This setup emphasizes a Python-based runtime (no separate Node/NPM installation required).
• The repository includes self-contained installers for other distribution methods, but the developer-focused path uses Python as shown above.

Tips and considerations:

• Security: Carefully manage which tools are exposed to AI agents and enable per-tool user approvals as needed.
• Docker sandboxing: Use the Docker-based sandboxing option to isolate tool execution when handling untrusted inputs.
• Endpoint configuration: The SSE URL is central to agent communication; ensure it is reachable by your MCP clients and extensions.
• Logging: Enable verbose logs during setup to audit tool calls and agent activity.
• Extensions and clients: Pair with the MCP Link browser extension or direct MCP clients for agent tooling discovery and execution.
• Updates: When upgrading, review the changelog and verify tool compatibility with your MCP clients.