Google launches vibe coding in AI Studio for faster app development
Google has debuted a new “vibe coding” flow in AI Studio that turns a single prompt into a working app, with a refreshed App Gallery, brainstorming tips during builds and an Annotation Mode for on‑screen edits.
Google has unveiled a vibe coding flow in Google AI Studio that aims to take developers from a single prompt to a working, multimodal app within minutes.
The company said the update is a way to cut through setup friction by letting Gemini models wire together the right tools and APIs automatically.
In AI Studio, “vibe coding” describes a new experience where developers state an idea (for example, a Veo‑powered video tool or an image editor) and the system assembles the underlying components, models and connections on their behalf.
It is complemented by a revamped App Gallery for inspiration, a Brainstorming Loading Screen that surfaces context‑aware ideas while projects compile, and an Annotation Mode that lets users highlight interface elements and ask Gemini to modify them directly.
Key changes: from inspiration to iteration
- Revamped App Gallery: Google has created a richer, visual catalogue with instant previews and starter code that users can remix.
- Brainstorming Loading Screen: While an app is building, AI Studio surfaces context‑aware suggestions to help refine ideas.
- Annotation Mode: Users are able to select an on‑screen element (such as a button or card) and instruct Gemini to change its style or behaviour.
- Quota continuity: If free usage is exhausted, developers can add their own API key to keep working; AI Studio will switch back once the free tier is renewed.
In the run‑up to the launch, Google also refreshed AI Studio with a single Playground for Gemini, GenMedia, text‑to‑speech and Live models; a clearer rate‑limit view; Maps grounding; and improved key management and saved system instructions.
How this helps developers
The approach reduces context‑switching and early plumbing work, allowing teams to move faster from idea to prototype while still retaining control over prompts, keys and model selection.
By integrating Gemini more tightly into the editor workflow (and surrounding it with clearer usage telemetry), AI Studio seeks to lower the barrier for newcomers without removing the levers experienced developers expect.
Alongside AI Studio’s changes, Google recently highlighted agentic capabilities such as Gemini 2.5 Computer Use, which developers can access via AI Studio and Vertex AI for tasks that require real browser navigation rather than API calls.