Google brings real-time collaborative vibe design to Stitch

Designing software usually involves a slow, frustrating cycle of back-and-forth mockups. Google wants to turn that into a live conversation. At I/O, the company announced major updates to its design tool Stitch, bringing real-time collaborative generation to the canvas.

Designing out loud

The biggest change is live feedback. The Stitch Agent now streams iterations directly onto the screen as you interact with it. Instead of waiting for a full render, you can use text prompts or simply speak out loud, watching your layout reflow and adjust in real time.

You do not have to start from scratch, either. Stitch now lets you import existing codebases and design files so the agent can iterate on top of what you already have.

Google also smoothed out the pipeline for taking designs out of the sandbox. You can generate a shareable preview link via Google AI Studio, export screens directly into Google Antigravity to plug in backend logic, or publish them straight to the web through Netlify. The update rolled out globally starting today.

Why it matters

For solo builders and product managers, this makes building throwaway prototypes almost frictionless. You no longer need to wireframe before you build; you just talk your way into a working interface.

We think Google's smartest move here is connecting the dots downstream. AI canvas tools often suffer from creating dead-end code that designers cannot actually hand off. By linking Stitch directly to Antigravity and Netlify, Google is turning generative doodles into functional web apps in a few clicks.

The catch, as always, will be code maintainability. It looks fantastic for early-stage brainstorming, but turning streaming components into clean, production-grade systems remains a real hurdle.