AI tooling overview

Meta publishes a suite of AI tools that supports agentic workflows for Meta VR development. Meta VR CLI drives a device from your editor or terminal, agent skills teach agents a Meta VR workflow, and LLM resources let any AI tool read current documentation.

Each tool stands on its own. Pick the ones that fit how you already work.

Meta VR CLI (metavr) drives your Meta VR device from a terminal, or lets an AI coding agent drive it through its Model Context Protocol (MCP) server. Both front ends reach device management, app installs, logs, performance traces, documentation search, 3D asset search, and Meta Horizon Store publishing.

Start with the Meta VR CLI overview for how the two front ends relate, or go straight to Install Meta VR CLI to set it up and connect an agent.

A skill teaches an AI agent to carry out one Meta VR job, such as analyzing a Perfetto trace or getting a build ready for the Meta Horizon Store. It spells out the steps, the tools, and the checks the agent would otherwise have to guess at. Skills follow the open Agent Skills specification, so they work in any agent that supports it.

See Browse agent skills for the published catalog, grouped by category.

Meta publishes the Meta VR developer documentation in a form built for large language models (LLMs). These pages are plain text, with no navigation, so they avoid bloating an agent’s context with unrelated page furniture. Point any AI tool at them and you get current answers about building for Meta VR.

See LLM resources for agents for the documentation and API reference indexes for each build path.