wot-mcp

wot-mcp

Bridges Web of Things (WoT) devices to AI assistants via MCP, enabling discovery, monitoring, and control of IoT devices through natural language.

Category
访问服务器

README

WoT-MCP

WoT-MCP is a server application that exposes Web of Things (WoT) devices to AI assistants via the Model Context Protocol (MCP).

As AI agents become more sophisticated, their ability to interact with the real world remains limited by fragmented IoT protocols. WoT-MCP solves this by translating the standardized Web of Things model (Properties, Actions, Events) into MCP primitives (Resources and Tools).

This allows any MCP-compliant AI client (like Claude Desktop or LangChain agents) to natively discover, monitor, and control physical devices without needing custom code for each device.

Features

  • Protocol Translation: Converts WoT Properties, Actions, and Events into MCP Resources and Tools.
  • Two Tool Strategies:
    • explicit: Generates individual tools for every property and action (e.g., set_temperature, get_humidity). Best for small numbers of devices.
    • generic: Provides a fixed set of tools (list_devices, read_property, write_property, invoke_action) to manage any number of devices. Best for scalability.
  • Transport Modes: Supports both stdio and streamable-http.
  • Event Buffering: Captures WoT events and exposes them as MCP resources.
  • Docker Support: Ready-to-use Dockerfile for containerized deployment.

Installation

git clone https://github.com/macc-n/wot-mcp.git
cd wot-mcp
npm install
npm run build

Usage

Stdio (Local Clients)

To use WoT-MCP with local clients like Claude Desktop, you can configure them to spawn the server directly.

Claude Desktop Configuration: Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "wot": {
      "command": "npm",
      "args": [
        "--prefix",
        "<absolute-path>/wot-mcp",
        "start",
        "--",
        "--tool-strategy",
        "explicit",
        "--config",
        "<absolute-path>/things-config.json"
      ]
    }
  }
}

Note: Ensure you use absolute paths for both the script and the configuration file.

Streamable HTTP

To expose the MCP server over HTTP:

npm start -- --mode streamable-http --port 3000 --config things-config.json

Claude Desktop Configuration: Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "wot": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "http://<remote-ip>:<port>/mcp",
        "--allow-http"
      ]
    }
  }
}

Tool Strategies

Explicit Strategy (Default) Creates a unique tool for every capability:

  • WoT Property:
    • Creates a getter tool.
    • If the property is writable, creates also a setter tool.
  • WoT Action:
    • Create a single tool with the input schema derived directly from the WoT Action input schema.
  • WoT Event:
    • Exposes a subscriptable resource.
npm start -- --tool-strategy explicit --config things-config.json

Generic Strategy Uses 4 static tools to manage all devices:

  • list_devices: Return a JSON list of all devices and their capabilties.
  • read_property: Takes devide_id and property_name.
  • write_property: Takes devide_id, property_name, and value.
  • invoke_action: Takes devide_id, action_name, and optional params.

Note: WoT Events are managed as described before.

npm start -- --tool-strategy generic --config things-config.json

Configuration File

You must load things from a JSON configuration file. The file supports HTTP, CoAP, and MQTT devices.

// things-config.json
{
  "things": [
    {
        "protocol": "http",
        "url": "http://localhost:8080/httpthermostat"
    },
    {
        "protocol": "coap",
        "url": "coap://localhost:5683/coaplight"
    },
    {
        "protocol": "mqtt",
        "url": "mqtt://test.mosquitto.org/MqttSensor",
        "td": "/path/to/mqtt-td.json"
    }
  ]
}

Note: For mqtt devices, the td field is required and must point to a local file containing the Thing Description, as TD discovery is not supported over MQTT.

Examples

The wot-mcp-cli repository contains an interactive Command Line Interface (CLI) client for the WoT-MCP server, allowing you to inspect tools and interact with devices.

The wot-mcp-examples repository contains sample code for devices, configuration files, and clients.

For comprehensive documentation and further details, please consult the respective repositories.

Docker

Build the image:

docker build -t wot-mcp .

Run with a configuration file:

docker run --rm --network="host" \
  -v $(pwd)/things-config.json:/app/things-config.json \
  wot-mcp \
  --tool-strategy explicit \
  --config /app/things-config.json \
  --mode streamable-http \
  --port 3000

Note: Replace the path of the config file.

Known Limitations

This first release focuses on the core functionality of bridging Web of Things devices to MCP. It does not yet include advanced features such as authentication, security mechanisms, or a persistent storage layer.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选