Meta launches Muse Code, an AI agent for complex code bases

Meta launched Muse Code, a terminal coding agent currently in beta, capable of performing complete software engineering tasks in large repositories. The agent, powered by the Muse Spark coding model, can plan changes, write code, and validate results. For sufficiently large projects, Muse Code splits into sub-agents that work in parallel in isolated environments without touching the user's local code. Internal testing demonstrated that it successfully built six game features simultaneously without conflicts. The solution positions Meta as more competitive against rivals like OpenAI's Codex and Anthropic's Claude Code, particularly in terms of cost.