xrblocks

Prototype immersive AI and XR experiences with hand tracking and gestures on three.js.

455|91|Updated Sep 25, 2025
One-click install
npx skills add https://github.com/google/xrblocks --skill xrblocks
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: xrblocks
Source: https://github.com/google/xrblocks/tree/main/src
Command: npx skills add https://github.com/google/xrblocks --skill xrblocks

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires three.js, tensorflow, pytorch, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill provides a comprehensive toolkit for quickly building immersive AI and XR applications, streamlining the development process and enabling developers to focus on creating engaging experiences.

Core Features & Use Cases

  • Hand Tracking & Gestures: Integrate advanced hand tracking and custom gestures with TensorFlow Lite / PyTorch models.
  • Gesture Recognition: Detect various hand gestures like pinch, open-palm, fist, thumbs-up, point, and spread.
  • World Understanding: Implement depth sensing, geometry-aware physics, and object recognition with Gemini.
  • Agent Context: Utilize semantic trees, visible-object state, and spatial UI for interactive experiences.
  • Use Case: Develop an interactive XR game where players can manipulate virtual objects using their hands, with AI-driven interactions and spatial audio.

Quick Start

Create a new XR Blocks application by running xb.init(new xb.Options()) in your script.

Frequently Asked Questions about xrblocks

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prototype an immersive WebXR experience with hand tracking?

To prototype an immersive WebXR experience with hand tracking, initialize your application using `xb.init(new xb.Options())`. This cross-platform SDK integrates TensorFlow Lite or PyTorch models to detect gestures like pinch, fist, and open-palm directly in extended reality.

What is needed to build AI-driven XR applications with three.js?

Building AI-driven XR applications with three.js requires Chrome v136+ with WebXR support on Android XR. You also need dependencies like TensorFlow and PyTorch to implement advanced features such as hand tracking, gesture recognition, and agent context.

Can I use Gemini for object recognition and world understanding in WebXR?

Yes, you can use Gemini for object recognition and world understanding in WebXR. The SDK facilitates depth sensing, geometry-aware physics, and object recognition, allowing developers to create interactive environments where AI responds to physical surroundings.

Does this SDK support spatial UI and semantic trees for agent context?

Yes, the SDK supports spatial UI and semantic trees for agent context. It enables interactive experiences by utilizing visible-object state and spatial audio, allowing AI agents to understand and react to the user's immersive environment.

How do I start building a cross-platform XR game with custom gestures?

Start building a cross-platform XR game by running `xb.init(new xb.Options())` in your script. You can then integrate custom gestures via TensorFlow Lite or PyTorch, enabling players to manipulate virtual objects using hand movements and AI-driven interactions.

What are the limitations of developing XR experiences for Android XR?

A key limitation of developing XR experiences is the strict environment requirement: it is designed specifically for Chrome v136+ with WebXR support on Android XR. Additionally, implementing specific features like gesture recognition requires external dependencies like three.js and TensorFlow.