Meta has quietly positioned itself as a latecomer in the AI assistant race, but a new release signals a more aggressive push. The company this week launched Muse Code, a terminal-based coding agent designed to help programmers handle complex tasks across large software repositories. Available in beta, Muse Code can plan changes, write code, and validate results, according to Meta CEO Mark Zuckerberg. The tool, installable with a single command, is powered by Meta's earlier Muse Spark coding model. For big jobs, it dispatches sub-agents that work in parallel within isolated worktrees, so the developer's main working copy remains untouched. Main Developments Muse Code's core strength lies in its ability to scale across entire repositories. Zuckerberg noted that during testing, the agent built six features for a game simultaneously without collisions, showcasing its parallel processing capabilities. Read also: DOJ oversight: OpenAI and Statsig settle green-card hiring claims The launch is part of Meta's broader effort to close the gap with AI lab competitors like OpenAI's Codex and Anthropic's Claude Code. Alexandr Wang, Meta's AI chief and head of Meta Superintelligence Labs, told the Wall Street Journal that the company believes Muse Code can be “an incredibly good option” for many workflows, particularly from a cost perspective. Background Meta has historically focused its AI investments on advertising, but the company has been diversifying. In June, it entered the enterprise AI market with a customer-service agent, marking a shift beyond its core ad business. Muse Code builds on this trajectory, leveraging the previously released Muse Spark model to extend Meta's AI reach into developer tools. The move also underscores Meta's increased spending on AI development, as it seeks to compete with well-funded rivals that have already established coding agents. Why It Matters For developers managing large code bases, Muse Code offers an affordable alternative to existing agents, potentially lowering the barrier to AI-assisted coding. Meta's entry into this space intensifies competition, which could drive down prices and spur innovation across the industry. For Meta, the release is a strategic step toward becoming a serious player in AI tools, not just a social media giant. It also signals a willingness to challenge OpenAI and Anthropic on their home turf, with a cost-focused pitch that could appeal to startups and enterprises alike. What's Next Muse Code is currently in beta, so developers can expect iterative improvements as Meta gathers feedback. The company has not announced a general availability timeline, but the focus on cost and performance suggests it will refine the agent's efficiency and expand its capabilities. Observers will be watching whether Muse Code can gain traction against entrenched competitors and whether Meta will integrate it into its broader enterprise AI offerings.