Google launches Playground AI platform for text-to-game creation

Google Labs has introduced Playground, a browser-based experimental platform that allows users to build and publish video games using natural-language text prompts.

The new platform, launched Wednesday, enables users to create 2D or 3D games without coding knowledge. According to TechCrunch, creators can select specific genres like racing or trivia, define gameplay mechanics, and upload visuals that the AI transforms into game assets. The tool uses a conversational interface to let users adjust physics, rules, and character designs by typing requests.

Playground is currently available to users aged 18 and older in the United States. While browsing and playing games is free, generating them relies on a weekly token system. The Verge reported that Google One AI subscribers will receive higher usage limits, while the underlying technology uses a combination of Google’s Gemini, Nano Banana, and Lyria models.

In partnership with Unity, Google also announced Unity Spark, a more advanced AI development engine. The Verge noted that Unity Spark will enter a closed beta later this year and will eventually integrate with Playground to provide professional-level mechanics. Games created on the platform can be kept private, shared via links, or published to an Explore gallery featuring leaderboards and multiplayer support.

The launch follows similar moves by competitors in the AI gaming space. TechCrunch pointed out that Roblox announced a natural-language build feature in July, while The Verge highlighted a recent Meta announcement regarding AI game-creation tools for its Horizon platform.

Background

Google has previously explored the gaming sector through its cloud-based Stadia service in 2022 and the more recent launch of YouTube Playables, a hub for instant web-based games.