Google AI Studio builds apps
Google AI Studio can now auto‑generate full‑stack apps — including DB, auth and deploy — from prompts, slashing build time for prototypes and internal tools (x.com). That could upend early-stage product discovery by letting founders iterate product‑market fit without deep engineering resources (x.com).
Google says the new build-mode agent is called “Antigravity” and the update is billed as a full‑stack “vibe coding” experience in a blog post by Ammaar Reshi and Kat Kampf. (blog.google) After user approval the agent will automatically provision Cloud Firestore for storage and Firebase Authentication for sign‑in flows, and it surfaces those choices in the Build settings. (blog.google) Projects can include a server‑side Node.js runtime so generated apps can run secure API calls, access third‑party services, and keep secrets out of client code. (ai.google.dev) The Antigravity agent can detect when libraries are needed, install npm packages on demand (examples cited include Framer Motion and shadcn), and update multiple files while preserving project context. (blog.google) Google says the system supports Next.js in addition to React and Angular, and the agent can create real‑time multiplayer features like collaborative workspaces or games. (blog.google) Build mode ties into Gemini capabilities — docs reference Nano Banana, the Live API, and selectable AI chips — and the Studio environment shows a live preview plus a Code tab for all generated files. (ai.google.dev)
