UnifiyMe
A Model Context Protocol (MCP) server for UniFi Network Controllers that enables LLMs to interact with UniFi infrastructure for monitoring and control.
README
UnifiyMe
A Model Context Protocol (MCP) server for UniFi Network Controllers. This server allows you to interact with your UniFi network infrastructure through LLMs, providing both high-level monitoring and deep system-level control.
Features
- Network Monitoring: Quick overview of devices, clients, and site health.
- Client Management: Block/unblock clients, force reconnections, and search historical connection data.
- Physical Control: Power cycle PoE ports and flash device LEDs for easy location.
- Advanced Analytics: Deep Packet Inspection (DPI) stats, subsystem health, and alarm tracking.
- Guest Access: Create guest WiFi vouchers on the fly.
- SSH Diagnostics: Execute commands directly on the router and fetch system logs for deep troubleshooting.
- Discord Integration: Manage and monitor your network directly from Discord.
- Prometheus Exporter: Export UniFi metrics to Prometheus for long-term monitoring and alerting.
- Grafana Dashboards: Pre-configured dashboards for visualizing network health and performance.
- Restricted YouTube Blocker: Specialized tools for managing YouTube access via UniFi firewall rules.
Optimization
This server includes specialized optimizations for UniFi controllers:
- Monkey-patched Session Management: Reduces redundant heartbeat checks, cutting request overhead by up to 50%.
- Intelligent Caching: 10-second TTL cache for frequent lookups to improve responsiveness and reduce controller load.
Reliability & Security
- Persistence Layer: Automated background monitoring ensures that critical traffic rules (like YouTube blocking) are re-enforced every 60 seconds if they are tampered with.
- Health Monitoring: Integrated
/healthendpoint and Docker health checks ensure maximum uptime and automated container recovery. - API Hardening: All Gateway actions are strictly validated with MAC address regex and required field checks.
- Secure Reverse Proxy: Pre-configured Caddy integration for automated TLS/HTTPS termination.
Setup
-
Clone the repository.
-
Install dependencies:
npm install -
Create a
.envfile with your UniFi credentials:UNIFI_HOST=https://your-unifi-controller-ip UNIFI_USERNAME=your-username UNIFI_PASSWORD=your-password UNIFI_SITE=default # Optional: SSH Credentials for diagnostics SSH_HOST=192.168.1.1 SSH_USERNAME=root SSH_PASSWORD=your-ssh-password # Optional: Discord Bot Configuration DISCORD_TOKEN=your-discord-bot-token DISCORD_CLIENT_ID=your-discord-client-id -
Build the project:
npm run build
Usage
1. Installation via npm (Published)
This is the recommended way for most users. You can run the server directly using npx in your Claude Desktop configuration.
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"unifyme": {
"command": "npx",
"args": ["-y", "unifiyme"],
"env": {
"UNIFI_HOST": "https://<CONTROLLER-IP>",
"UNIFI_USERNAME": "<YOUR-USERNAME>",
"UNIFI_PASSWORD": "<YOUR-PASSWORD>",
"UNIFI_SITE": "default"
}
}
}
}
2. Local Development Setup
If you prefer to run from source or are contributing:
- Clone the repository.
- Install dependencies:
npm install - Build the project:
npm run build - Use the following config:
{
"mcpServers": {
"unifyme": {
"command": "node",
"args": ["/path/to/UnifiyMe/dist/index.js"],
"env": {
"UNIFI_HOST": "...",
"UNIFI_USERNAME": "...",
"UNIFI_PASSWORD": "...",
"UNIFI_SITE": "default"
}
}
}
}
Discord Bot
To start the Discord bot:
npm run start # If built
# OR
npm run dev # For development
Prometheus Exporter
The server automatically starts a Prometheus exporter on port 9090 (configurable). Metrics are available at /metrics.
Docker Support
You can run the entire stack (UnifiyMe, Prometheus, Grafana) using Docker Compose:
docker-compose up -d
Tools
Monitoring & Discovery
get_network_status: Get overall network health and status.list_devices: List all network devices (APs, switches, gateways).list_clients: List all connected network clients.get_device_details: Get detailed information about a specific device.search_network: Search for devices or clients by IP, MAC, hostname, or alias.get_bandwidth_stats: Get top bandwidth consumers.get_network_topology: Get network topology (simplified view of uplinks).get_client_history: Search historical client data.get_network_health: Get detailed subsystem health status.get_alarms: List recent alerts and IPS/IDS events.get_dpi_stats: Get application usage statistics.get_wlan_config: Get WiFi network configurations (SSIDs, security, etc.).get_network_config: Get logical network configurations (VLANs, subnets, DHCP).get_firmware_status: Check for available firmware updates across all devices.get_firewall_rules: Get all configured firewall rules.get_firewall_groups: Get all configured firewall groups (IP, Port).get_deep_dive: Perform a deep dive into network health, DPI stats, and active client traffic.get_events: Get recent network events.
Control & Management
reboot_device: Restart a device by MAC address.cycle_poe_port: Power cycle a PoE port on a switch.locate_device: Flash/Stop flashing the LED on a device.block_client/unblock_client: Manage network access for clients.reconnect_client: Force a client to reconnect.create_voucher: Create guest WiFi vouchers.set_client_user_group: Assign a client to a specific user group (for throttling).create_user_group: Create a new user group with bandwidth limits.
Specialized Management
block_restricted_youtube/unblock_restricted_youtube: Manage YouTube access for Restricted devices.get_restricted_youtube_status: Check current YouTube blocking status for Restricted.detect_iot_devices: Identify potential IoT devices not on the IoT VLAN.migrate_iot_devices: Propose migration of detected IoT devices to a target network.enforce_iot_limits: Throttle high-bandwidth IoT devices.protect_vips: Ensure VIP devices are not throttled.ensure_traffic_rule: Create or update complex traffic rules.
Diagnostics (SSH)
run_ssh_command: Execute a command directly on the router.get_router_logs: Fetch system logs from the router.
Contributing
Contributions are welcome! Please see CONTRIBUTING.md for guidelines on how to report issues, suggest features, and submit pull requests.
License
This project is licensed under the MIT License.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。