MCP Server

Note

This feature is only available with a paid license.

The Model Context Protocol (MCP) is a standard that allows AI assistants (like Claude) to connect to your applications.

VRTI includes a built-in MCP server that allows you to give AI assistants direct access to your treadmill. This means you can ask your AI assistant to check your stats, change settings, or even control the treadmill speed for you, all through natural language.

Tip

The MCP Server automatically provides the AI with detailed instructions on how VRTI works, so you don’t need to explain concepts like “Speed Multipliers” or “AutoWalk” to it manually.

Configuration

The MCP Server can be enabled within VRTI’s settings.

1. Open VRTI Settings.
2. Navigate to the MCP Server section.
3. Toggle Enable MCP Server to On.

By default, the server runs on port 47127 at the following address:

http://localhost:47127/sse

Note

This is an SSE (Server-Sent Events) endpoint. When configuring your MCP client, ensure you use the HTTP/SSE connection type, not the command/stdio type.

Connecting Clients

You can connect any MCP-compliant client to VRTI. Below are instructions for some popular clients.

Claude Desktop

To use VRTI with the Claude Desktop app, you need to edit your configuration file.

Prerequisite

You need to have Node.js installed on your computer to use the npx command.

1. Locate your claude_desktop_config.json file:
   • Windows: %APPDATA%\Claude\claude_desktop_config.json
   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
2. Add the VRTI server configuration:

{
  "mcpServers": {
    "VRTI": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "http://localhost:47127/sse",
        "--allow-http"
      ]
    }
  }
}

3. Restart Claude Desktop. You should see VRTI as an available MCP server in the Claude Desktop app’s settings, in the “Connectors” section.

Cursor

To use VRTI with Cursor:

Prerequisite

You need to have Node.js installed on your computer to use the npx command.

1. Open File > Preferences > Cursor Settings
2. Navigate to Tools & MCP.
3. Click + New MCP Server.
4. Add the VRTI configuration to your mcp.json file:

{
  "mcpServers": {
    "vrti": {
      "url": "http://localhost:47127/sse",
      "transport": "sse"
    }
  }
}

5. Save your mcp.json file. You can also find this file in the $HOME/.cursor/mcp.json directory.
6. Restart Cursor and return to the MCP settings. You should see VRTI in the list with a green indicator.

gemini-cli

To use VRTI with gemini-cli:

1. Install gemini-cli if you haven’t already.
2. Open your gemini-cli configuration file (usually ~/.gemini/settings.json).
3. Add VRTI to the mcpServers configuration:

{
  "mcpServers": {
    "vrti": {
      "httpUrl": "http://localhost:47127/sse"
    }
  }
}

4. Save the file. You can now use VRTI tools within gemini-cli.

Capabilities

Once connected, the AI assistant has access to a wide range of tools and resources.

Controlling the Treadmill

The AI can actively control your treadmill and VR settings:

• Set Target Speed: “Set the treadmill to 5 km/h”
• Toggle AutoWalk: “Turn on AutoWalk”
• Speed Adjustments: “Set speed multiplier to 1.5” or “Limit my speed to 10 km/h”
• Reset Stats: “Reset my session statistics”

Monitoring

The AI can read the current state of your session:

• Current Status: “Is the treadmill running?” or “How fast am I going?”
• Statistics: “How far have I walked today?” or “Show me my monthly stats”
• Settings: “What are my current AutoWalk settings?”

Available Tools & Resources

VRTI exposes the following tools and resources to connected MCP clients.

Tools

Tools allow the AI to perform actions or fetch specific data on demand.

Tool Name	Description
vrti_set_target_speed	Set the target speed of the treadmill in km/h.
vrti_get_treadmill_state	Get the current status (running, speed, connection) of the treadmill.
vrti_toggle_autowalk	Enable or disable the auto walk feature.
vrti_get_autowalk_settings	Get the current configuration for the auto walk feature.
vrti_set_speed_multiplier	Set the auto walk speed multiplier (e.g., 1.5 for 50% faster).
vrti_set_speed_offset	Apply a temporary speed offset (e.g., +0.5 km/h) to the auto walk speed.
vrti_set_speed_override	Force the auto walk speed to a fixed value (or -1 to disable).
vrti_set_speed_limit	Set the maximum speed limit for the user.
vrti_center_yaw_lock	Recenter the Yaw Lock direction.
vrti_toggle_yaw_lock	Enable or disable Yaw Lock.
vrti_set_input_mode	Change the input mode (e.g., to VRChatOSC or XINPUT).
vrti_get_statistics	Get statistics for the current ongoing session.
vrti_reset_statistics	Reset the current session’s statistics.
vrti_get_historic_daily_stats	Retrieve historical statistics grouped by day.
vrti_get_historic_monthly_stats	Retrieve historical statistics grouped by month.
vrti_get_historic_sessions	Retrieve a list of past individual sessions.

Resources

Resources provide direct access to data that can be read by the AI to understand the current context.

Resource URI	Description
vrti://treadmill/state	JSON object containing the real-time state of the treadmill (speed, connection, etc.).
vrti://autowalk/settings	JSON object with the current auto walk settings.
vrti://autowalk/speed	JSON object with the current auto walk speed being sent to your game.
vrti://statistics/current	JSON object with live statistics for the current session.