Meta has released Muse Code, a terminal-based coding agent designed to complete complex software engineering tasks across large code repositories. The tool is currently available in beta and is powered by Meta’s Muse Spark 1.2 coding model.
Meta CEO Mark Zuckerberg said Muse Code can plan software changes, write code and validate the results. It is intended to handle larger engineering projects by splitting work across multiple AI agents that can operate simultaneously.
Muse Code Can Run Multiple Agents in Parallel
For complex tasks, Muse Code can create separate sub-agents and assign each one a different part of the job. Those agents work in isolated Git worktrees, allowing them to make changes without interfering with a developer’s existing working copy.
Zuckerberg said Meta tested the system by having it build six features for a game at the same time without creating conflicts between the agents. The approach is designed to make the tool more useful for large repositories where several related changes may need to happen together.
Meta provides installation through a single terminal command. The company outlined the agent and its underlying Muse Spark 1.2 model in an official developer announcement.
Meta Is Competing on Price
Muse Code puts Meta more directly against coding agents such as OpenAI’s Codex and Anthropic’s Claude Code. Meta is positioning price as one way to attract developers to its offering.
Alexandr Wang, who leads Meta Superintelligence Labs, told The Wall Street Journal that Muse Code could be particularly competitive for workflows where cost is important. Meta is offering two pricing options, including a discounted programme priced at 20 cents per million output tokens for users who provide feedback.
The coding agent is another product from Meta’s reorganised AI operation under Wang. Meta hired the former Scale AI chief executive in 2025 as part of a broader effort to improve its position against companies including OpenAI, Anthropic and Google.
Meta Expands Beyond Consumer AI
Meta has been increasing its AI spending while developing products beyond the recommendation and advertising systems that underpin Facebook and Instagram.
In June, the company moved further into enterprise AI with a business agent designed to automate customer support and other routine company operations. Muse Code extends that strategy into software development, where AI coding agents have become one of the most actively contested areas of the market.
Meta is also continuing to invest heavily in computing infrastructure and AI talent. Muse Code gives the company another way to turn that spending into a developer-facing product while testing whether its Muse models can gain adoption outside Meta’s own platforms.
Featured image credits: TwoTone Creative
For more stories like it, click the +Follow button at the top of this page to follow us.