MCP Server UniFi

Name: MCP Server UniFi
Author: zcking

@zcking

Repository

About MCP Server UniFi

mcp-server-unifi

#mcp-server-unifi #network-management #ai-integration

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "mcp-server-unifi": {
      "command": "uv",
      "args": [
        "venv"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is MCP Server UniFi?

MCP Server UniFi is a server that wraps the UniFi Network API, enabling natural language interactions with your UniFi network devices through AI agents like Goose and Claude. It allows querying sites and devices on a local UniFi network using the Management Control Protocol (MCP).

How to use MCP Server UniFi?

Obtain a UniFi API key from your UniFi console (Settings » Control Plane » Integrations). Clone the repository, create a virtual environment with uv, install dependencies via uv sync, and set the UNIFI_API_KEY environment variable. Start the development server with uv run mcp dev main.py, then integrate with Goose AI (via custom extension configuration) or Claude Desktop (by editing claude_desktop_config.json).

Key features of MCP Server UniFi

Query UniFi sites and devices using natural language
Works with both Goose AI and Claude Desktop
Local server connects directly to your UniFi Network application
Secure API key‑based authentication

Use cases of MCP Server UniFi

Ask an AI agent, “How many UniFi devices are on my network?”
Get detailed status of a specific UniFi site
Monitor device connectivity without logging into the UniFi console

FAQ from MCP Server UniFi

What prerequisites are needed to run this server?

Python 3.8 or higher, the uv package manager, a UniFi Network application, and a UniFi API key.

How do I obtain a UniFi API key?

Go to your UniFi console at https://unifi.ui.com, navigate to Settings » Control Plane » Integrations, and click “Create API Key”.

Which AI agents are compatible with this server?

Goose AI and Claude Desktop are explicitly supported out of the box.

How do I start the MCP server for testing?

Run uv run mcp dev main.py from the project directory. The MCP Inspector will be available at http://localhost:5173.

How do I configure Claude Desktop to use the server?

Open Claude Settings » Developer » Edit Config, and add an entry to claude_desktop_config.json specifying the command uv with arguments --directory /path/to/mcp-server-unifi run main.py.