MCP Server introduces a new way to interact with OrbitOS devices — built for automation, AI, and modern developer workflows.
- The first version is already available in the store and ready to install remotely on your devices.
Here
![[Image: 9TdYBjZ.png]](https://i.imgur.com/9TdYBjZ.png)
It exposes the device through the Model Context Protocol (MCP), enabling tools like Cursor and other compatible AI assistants to interact with the system using structured, well-defined capabilities — no custom scripts, no manual glue code.
- What does that actually mean?
Instead of writing scripts or manually connecting to your device, interactions become simple and direct.
With an MCP-compatible tool, you can:
- What you can do with it
Inspect the system
- Example: Natural Language Interaction
Once connected to an MCP compatible client, you can interact with your device (e.g. Raspberry Pi or Arduino UNO Q) using simple natural language sentences such as:
- Get device info and resources usage
![[Image: gpb9ywF.png]](https://i.imgur.com/gpb9ywF.png)
- List GPIOS
![[Image: uAARk2y.png]](https://i.imgur.com/uAARk2y.png)
- List installed applications
![[Image: EpxGt7d.png]](https://i.imgur.com/EpxGt7d.png)
- Turn GPIO17 to hight level
![[Image: FfbAWLP.png]](https://i.imgur.com/FfbAWLP.png)
⚙️ Client Configuration (ex. Cursor config)
- Final note
MCP Server moves OrbitOS devices closer to AI-native infrastructure.
Define capabilities once — and any MCP-compatible tool can use them.
- The first version is already available in the store and ready to install remotely on your devices.
Here
![[Image: 9TdYBjZ.png]](https://i.imgur.com/9TdYBjZ.png)
It exposes the device through the Model Context Protocol (MCP), enabling tools like Cursor and other compatible AI assistants to interact with the system using structured, well-defined capabilities — no custom scripts, no manual glue code.
- What does that actually mean?
Instead of writing scripts or manually connecting to your device, interactions become simple and direct.
With an MCP-compatible tool, you can:
- Query system status
- Install packages
- Reboot the device
- Read or control GPIO
- What you can do with it
Inspect the system
- Hardware info, OS details, metrics, API versions
- List installed apps
- Install packages
- Remove packages
- Install OTA updates
- Perform factory reset
- Reboot or shut down with context
- Read/write pins
- Change direction
- Access by name or ID
- Example: Natural Language Interaction
Once connected to an MCP compatible client, you can interact with your device (e.g. Raspberry Pi or Arduino UNO Q) using simple natural language sentences such as:
- "What’s the current status of the device?"
- "List installed applications"
- "Install this package from file"
- "Reboot the device"
- "Set GPIO 17 to HIGH"
- "Check system metrics"
- Get device info and resources usage
![[Image: gpb9ywF.png]](https://i.imgur.com/gpb9ywF.png)
- List GPIOS
![[Image: uAARk2y.png]](https://i.imgur.com/uAARk2y.png)
- List installed applications
![[Image: EpxGt7d.png]](https://i.imgur.com/EpxGt7d.png)
- Turn GPIO17 to hight level
![[Image: FfbAWLP.png]](https://i.imgur.com/FfbAWLP.png)
⚙️ Client Configuration (ex. Cursor config)
Code:
{
"mcpServers": {
"orbit-os": {
"url": "http://<DEVICE_IP>:9999/mcp"
}
}
}- Final note
MCP Server moves OrbitOS devices closer to AI-native infrastructure.
Define capabilities once — and any MCP-compatible tool can use them.

