Develop
Develop
Select your platform

AI tooling overview

Updated: Feb 27, 2026
AI-powered tooling is transforming how developers build for Meta Horizon OS. By integrating large language models (LLMs) directly into your development workflow, you can query documentation instantly, debug device issues in real time, search Meta’s 3D asset library with natural language, and analyze performance traces all without leaving your editor.

Why accelerate with AI?

Building for spatial computing involves working across multiple SDKs, device-specific debugging workflows, and performance-sensitive rendering pipelines. Traditional development cycles require context-switching between documentation portals, ADB commands, trace analysis tools, and asset pipelines.
AI tooling collapses these steps. Instead of navigating docs manually, you describe what you need. Instead of writing ADB commands from memory, you ask. Instead of browsing asset catalogs, you search with plain language. The result is less friction, faster iteration, and more time spent building.

Use cases

  • Documentation retrieval — Query Unity, Unreal Engine, Meta Spatial SDK, Android, Native C++, and Web development docs for Meta Quest through natural language.
  • Device debugging — Pull logcat logs, stream real-time device output, capture screenshots, and filter by tag, level, package, or PID on connected Meta Quest devices.
  • Performance analysis — Load Perfetto traces, query thread state, run SQL against trace data, and extract GPU frame metrics for profiling sessions.
  • 3D asset discovery — Search Meta’s asset library for existing 3D models using text descriptions to accelerate prototyping and content creation.
  • Timestamp conversion — Convert hexadecimal timestamps from trace data to human-readable dates and times. This is useful for correlating events across different logs and traces.

Agentic tools and skills

The meta-quest/agentic-tools⁠ GitHub repository provides a collection of ready-made skills that onboard your AI agent to build for Meta Horizon OS. These skills contain distilled knowledge from Meta’s developer documentation and from engineers across DPP who have solved common problems developers encounter when building for the platform — from converting Android apps to Horizon OS to diagnosing jank with Perfetto.
The repository also bootstraps the Horizon Debug Bridge (hzdb) CLI and MCP server, so your agent has access to documentation, device debugging, and asset search tools out of the box.

Compatibility

The skills are packaged in a format that works with more than a dozen AI developer tools, including:
  • Claude Code
  • Cursor
  • GitHub Copilot
  • Gemini CLI
  • Augment Code
  • Roo Code
  • And many more

Install

You can add the skills to your project with a single command using your tool’s built-in skills support, or through skills managers like npx skills from Vercel. See the repository README⁠ for setup instructions specific to your tool.
Did you find this page helpful?
Nav Logo
Virtual reality / Meta Horizon
Virtual reality / Meta Horizon
Virtual reality / Meta Horizon
English (US)
© 2026 Meta