RegenNexus

You deploy RegenNexus MCP Server to expose RegenNexus hardware capabilities to MCP clients, enabling AI-driven control of GPIO, PWM, robotics, sensors, and cameras. It supports local (direct import) and remote (HTTP API) operation, auto-discovery of nodes, and mesh networking, so you can build responsive, hardware-aware AI workflows with Claude Code, Claude Desktop, and other MCP clients.

How to use

You run the MCP server on the machine where RegenNexus is installed. Choose Local Mode for minimal latency or Remote Mode to connect to a UAP service over HTTP.

In Local Mode, load the MCP server directly on the same machine as RegenNexus. In Remote Mode, your MCP client talks to RegenNexus through an HTTP API exposed by a UAP service. Auto Mode tries Local first and automatically falls back to Remote if Local is not available.

From your MCP client, you access a set of tools that map to RegenNexus hardware and network features. Typical workflows include listing devices, querying sensor data, adjusting GPIO or PWM, moving robotic arms, and communicating with mesh nodes.

How to install

Prerequisites: Python 3.8+ and pip must be available on your system.

pip install regennexus-mcp[local]

This installs the MCP server with local (UAP import) support and minimal dependencies for direct hardware control.

If you prefer remote API support, install the remote-enabled package variant and run the server with the appropriate mode.

pip install regennexus-mcp[remote]

Run the MCP server in Local Mode to start immediately against local UAP hardware.

REGENNEXUS_MODE=local regennexus-mcp

Or run in Remote Mode after you configure the endpoint and API key on the MCP client side.

REGENNEXUS_MODE=remote
REGENNEXUS_ENDPOINT=http://uap-server:8080
REGENNEXUS_API_KEY=your-api-key
regennexus-mcp

## Configuration and usage notes

Configure your MCP client to connect to RegenNexus using a stdio command or via HTTP, depending on the mode you choose. The local mode uses direct imports for the fastest interaction, while remote mode relies on the UAP HTTP API.

Environment variables you may set include the mode, endpoint, API key, and log level. Use these to tailor behavior for debugging or production.

## Available tools

### gpio\_write

Set a GPIO pin to HIGH (1) or LOW (0)

### gpio\_read

Read the current state of a GPIO pin

### pwm\_write

Set PWM duty cycle (0-100%) for motors, LEDs, servos

### read\_sensor

Read value from a sensor (temperature, humidity, etc.)

### i2c\_scan

Scan I2C bus for connected devices

### serial\_send

Send data over serial port (UART)

### serial\_read

Read data from serial port

### robot\_arm\_move

Move a robotic arm to specified joint positions

### gripper\_control

Open or close a robotic gripper

### list\_devices

List all connected hardware devices

### device\_info

Get device details (CPU, memory, IP, temperature)

### camera\_capture

Capture a single image from a camera

### list\_nodes

List all nodes in the mesh network

### ping\_node

Ping a node and measure network latency

### send\_to\_node

Send a message/command to a specific node

### broadcast\_message

Broadcast a message to all nodes

### find\_by\_capability

Find nodes with a specific capability