Google revamps AI Studio with new features for vibe coding

Google is rolling out a revamped user experience for AI Studio, positioning it as a new entry point for building AI-powered applications with less friction from prompt to production. This update targets developers, startups, and even non-coders interested in rapid AI prototyping, reflecting Google’s aim to democratize app creation.

The new experience was first hinted at by Logan Kilpatrick as a way to lower the technical barrier, and it’s only the initial stage, with more upgrades expected within two months. Many anticipate these will be boosted further by the release of Gemini 3, which is expected to offer advanced coding capabilities.

The feature set now includes an application gallery, which may evolve to showcase user-created apps, resembling the artifact galleries on platforms like Claude. The build section now starts with a refreshed page featuring a new model selector, currently supporting pro and flash variants, with room for future additions. Importantly, secret variable support will allow developers to safely store sensitive keys when building and publishing apps, a key requirement for production use.

A notable addition is a grid of AI features that can be added to prompts with a single click. These “superpowers” include tools for editing media, accelerating AI responses, or requesting deeper reasoning from Gemini. The modular approach enables users to mix capabilities that will be recognized as components by Gemini during app generation, streamlining both user intent and technical implementation.

The interface also introduces a playful “I’m Feeling Lucky” button, offering random prompt suggestions to spark new ideas. In the app editor, context-aware suggestions from Gemini guide feature additions, like proposing an AI-powered functionality or UI improvements. Users can now also annotate specific UI parts and instruct Gemini to adjust them, refining customization.

Apps can be easily deployed to Google Cloud Run via a single click, granting a live URL for immediate access. While the specifics of app publishing, whether to a public gallery or elsewhere, remain unconfirmed, this step points toward broader discoverability and sharing in the future. Google’s larger product vision is to make AI-powered development accessible at scale, aiming for one million apps built by year’s end. These changes align with Google’s broader push into agentic AI, making AI Studio central to its ecosystem and reinforcing its competitive stance against rivals like Anthropic and OpenAI.