MCP API Bridge Server

You can bridge popular data sources and APIs using the MCP API Bridge Server. It connects Google Sheets, Azure AI, and MQTT ecosystems through a unified MCP interface, letting you generate code templates, run AI-assisted tasks, and manage IoT devices from a single, consistent client experience.

How to use

You use an MCP client to call predefined tools that generate code snippets, manage AI conversations, and handle MQTT-based IoT workflows. Start by ensuring your MCP client can reach the server endpoints, then select a tool by category (Google Sheets API, Azure AI API, or MQTT API). Each tool returns ready-to-adapt code or actions you can execute in your project. You can combine tools to create end-to-end data flows, such as collecting sensor data via MQTT, storing it in Google Sheets, and analyzing it with Azure AI.

To begin experiments, install the MCP API Bridge Server globally, then connect your MCP client to the server and list available tools. Use the provided tool names to call specific actions, such as generating Google Sheets code templates or initiating AI-driven analysis, and adapt the generated code to your environment.

How to install

Prerequisites you need before installation:

• Node.js 18.0.0 or newer

• npm or yarn

• Cursor IDE (optional, if you intend to use Cursor)

Follow these steps to get started quickly

1. Install the MCP API Bridge globally

npm install -g https://github.com/marty5499/mcp-api-bridge.git

2. If you use Cursor, configure MCP in your Cursor settings

# macOS
~/.cursor/mcp.json

# Linux
~/.config/cursor/mcp.json

# Windows
%APPDATA%\Cursor\mcp.json

Add the MCP server entry to Cursor’s config so it can launch the bridge

{
  "mcpServers": {
    "api-bridge": {
      "command": "mcp-api-bridge",
      "env": {}
    }
  }
}

3. Restart Cursor to apply the new MCP server connection

4. Update to the latest version when needed

npm update -g https://github.com/marty5499/mcp-api-bridge.git

5. For local development and testing, you can install dependencies and run tests

# Install dependencies
npm install

# Run basic tests via provided usage example
# Example test command (from the source):
node mcp-api-bridge.js

# Start development mode with file watching
npm run dev

Configuration and endpoints

End points you can use with the MCP API Bridge Server are categorized by API type.

Google Sheets API base URL

https://hshgpt.webduino.tw/api/sheets/ (operations: append, get, update, del, save)

Azure AI API uses WebSocket communication

WebSocket URL: wss://hshgpt.webduino.tw

MQTT API broker URL and credentials

Broker URL: wss://mqtt-edu.webduino.io/mqtt Username: hsh2025 Password: hsh2025

Practical usage examples

The server provides concrete tool templates you can adapt in your projects. For instance, you can generate code to append rows to a Google Sheet, read sheet data, update or delete rows, or overwrite the sheet. You can also create an AI-enabled chat workflow with Azure AI and manage IoT devices and messages through MQTT.

Example workflows you can build include IoT data collection pipelines and AI-assisted data analysis. Create a temperature sensor device, publish readings to a server, store the data in Google Sheets, and run Azure AI analysis on the sheet data.

Troubleshooting

If you encounter issues, check that the MCP server is reachable from your client, verify endpoint availability, and review server logs for errors. For Cursor users, ensure the MCP JSON configuration is correctly formatted and that Cursor has been restarted after configuration changes.

Available tools

google_sheet_append

Generate code to append new rows to a Google Sheet using the Sheets API.

google_sheet_get

Generate code to read data from a Google Sheet.

google_sheet_update

Generate code to update specific rows in a Google Sheet.

google_sheet_delete

Generate code to delete specific rows in a Google Sheet.

google_sheet_save

Generate code to overwrite an entire Google Sheet with new data.

azure_ai_chat

Generate code to interact with Azure AI (GPT-4o-mini) for chat-style conversations, with options for streaming and non-streaming modes.

mqtt_device_create

Create and register an IoT device for MQTT-based communication.

mqtt_publish

Publish a message to an MQTT topic (supports synchronous and asynchronous modes).

mqtt_publish_sync

Publish a message synchronously and wait for a response with a timeout.

mqtt_register_handler

Register a message handler for processing incoming MQTT messages.

mqtt_subscribe

Subscribe to an MQTT topic and receive messages according to QoS settings.