·3 min read·Product

Meta launches Muse Code, a coding agent built for very large repositories

It runs in the terminal, it is in beta and it is powered by Muse Spark. Its distinguishing feature: it spawns parallel sub-agents, each in an isolated environment, so several features can be built at once without colliding — and without touching your working copy.

  • What happened: Meta moves squarely onto Claude Code and Codex territory.
  • The differentiator: isolated parallel sub-agents; Meta cites "six game features simultaneously with no collisions".

Muse Code, announced on August 5, is Meta's coding agent for large code bases. It installs with a single command, runs from the terminal and is powered by the Muse Spark model. It is available in beta.

How it works: instead of working sequentially on your local copy, it spawns sub-agents that run in parallel in isolated environments. Meta's example is "building six features for a game simultaneously with no collisions". Your working copy is left untouched.

Why it matters: competition in coding agents is no longer about who writes the better function, but about who best orchestrates many tasks at once across a real repository. With Meta in, the three majors — Anthropic, OpenAI and now Meta — are competing in the same format: terminal, agent, repository.

Go to TechCrunch's coverage →

Source: TechCrunch

Go to the original source →

← Back to all news