Agents

Meta rolls out Pocket AI game-creation app in the US

Meta has launched Pocket, an experimental AI-powered game creation app, to US users, signaling a major push to make vibe-coding and rapid software development accessible to the public.

TechCrunch AI2 days agoAgents
Image: TechCrunch AI

Meta is expanding the release of Pocket, its experimental mobile application that allows users to build and share simple interactive games using natural language AI prompts. After a quiet initial test phase in Brazil last month, the app is now rolling out to all users in the United States. Pocket represents Meta's latest effort to bring consumer-facing AI creation tools into the mainstream, following the release of its image-generation features, the experimental video app Vibes, and the recently launched Meta AI app for Mac.

The technology behind Pocket stems from Meta's acquisition of the team behind Gizmo, a vibe-coded gaming platform developed by Atma Sciences. With the official launch of Pocket, Meta is shutting down the original Gizmo application. Inside Pocket, users generate interactive experiences—which Meta refers to as "gizmos"—that respond to physical phone tilts, touch inputs, sound effects, and music clips. These creations can also integrate a user's camera roll or live camera feed. Once published to a scrollable feed, other users can play, repost, or remix the games into new variations.

For software developers and product managers, Pocket demonstrates how major tech firms are leveraging AI to accelerate their own development pipelines. Meta CEO Mark Zuckerberg recently highlighted that AI-assisted coding has dramatically shortened the company's timeline for testing and shipping new software. This accelerated pipeline has recently produced several other standalone applications, including Instagram Instants, the Forum groups app, the Seller marketplace app, and an AI-driven bedtime story generator.

This shift toward vibe-coding tools suggests a future where the barrier to software creation is virtually non-existent for everyday consumers. For industry practitioners, it signals a transition where platforms must support highly modular, remixable code architectures. As Meta continues to scale these experimental applications using its internal recommendation algorithms, developers will need to adapt to an ecosystem where user-generated, AI-prompted micro-apps can be spun up and shared instantly.

This is our own summary of reporting by TechCrunch AI

More in Agents