Stitch 3.0: AI UI generation with direct Figma export

What it is. Stitch is a Google tool that generates UI screens for mobile and web from a text prompt. You type what you need, and it renders actual interface layouts on a live canvas. From there you can refine in place, stream edits, and export directly to Figma, Netlify, Lovable, or Bolt.

Why it matters. The gap between a brief and a first real screen is usually where time gets wasted, either in a blank Figma file or a back-and-forth with a developer. Stitch shortens that stretch. For a designer or product team trying to show a client something that looks like a real product, getting to a credible first screen in minutes rather than hours is worth paying attention to.

How to use it. Start with a prompt that describes the screen you actually need, the more specific the better. Something like “onboarding flow for a fitness app, three steps, progress indicator at the top” gives it more to work with than a vague category. The canvas updates as it generates, so you can interrupt and redirect if the first pass is off. Once the layout is close, use the in-place editing to adjust components or copy directly on the canvas. When it looks right, hit export and point it at Figma. The layers come through as editable elements, so a designer can pick up from there without having to redraw anything.

Limits. The outputs are starting points, and they will need design polish before they are ready for handoff to engineering. Generated layouts can be generic and component choices may not match your existing design system, which means time spent reconciling things in Figma anyway. Complex interactions and animations are outside what a prompt-to-screen tool like this can express, so anything beyond static layouts needs further work. It is also a Google Labs product, which means the roadmap and long-term availability are not guaranteed.