Turning a napkin sketch or a rough idea into a working app interface typically demands hours of manual design and frontend coding. Google’s new AI-powered tool, Stitch, eliminates those bottlenecks by translating natural language prompts or reference images directly into functional, export-ready UI designs. Announced at Google I/O 2025 and now available as an experiment in Google Labs, Stitch aims to streamline the journey from concept to code for both experienced developers and design newcomers.

Generating UI Layouts with AI-Powered Prompts

Stitch uses Google’s Gemini 2.5 Pro and Gemini 2.5 Flash models to interpret user input and quickly generate complex user interfaces. Developers can describe the desired app screen using plain English, specifying details like color themes, layout preferences, or even the overall user experience. For those who already have a visual direction, uploading wireframes, sketches, or screenshots allows Stitch to use these references as a blueprint, producing multiple interface variants in minutes.

The AI-driven workflow solves a persistent problem: bridging the gap between initial design ideas and production-ready code. Instead of painstakingly building each UI component and styling from scratch, users receive a full set of HTML and CSS code alongside the visual design. This code can be immediately exported and refined in an integrated development environment (IDE) or shared with team members for further iteration.

Exporting and Refining Designs in Figma

Stitch’s export options go beyond raw code. With one click, users can send generated designs directly to Figma, the industry-standard collaborative design platform. This makes it easy to polish visual details, tweak layouts, or collaborate with other designers before final implementation. The Figma export ensures that even if the initial AI-generated result isn’t perfect, teams can quickly iterate without starting from scratch.

For developers who want hands-on control, Stitch exposes all the generated frontend code. This transparency allows for deeper customization and seamless integration into existing projects, sidestepping the limitations of black-box AI tools that don’t provide editable output.

Iterative Editing and Upcoming Features

Stitch supports iterative design, letting users fine-tune elements after the initial generation. According to Google, an upcoming update will enable users to modify UI elements simply by taking a screenshot of the area they want to change and annotating it with their desired adjustments. This approach promises to make UI editing even more intuitive, especially for those less comfortable with traditional design tools.

While Stitch accelerates the early stages of UI creation, it’s not intended as a full replacement for established platforms like Figma or Adobe XD. Instead, it’s positioned as a rapid prototyping tool—ideal for quickly producing first drafts, experimenting with different looks, or getting stakeholder feedback before investing in manual refinement.

How Google Stitch Compares to Other AI Coding Tools

The market for AI-driven app development has grown rapidly, with tools like Cursor, Cognition’s Devin, and Microsoft’s GitHub Copilot offering code generation and agentic workflows. Stitch stands out by focusing specifically on UI design, combining visual generation with code output and built-in export to design platforms. Its support for both text and image prompts, plus the ability to generate multiple UI variants, gives it a unique edge for teams that need to iterate quickly on frontend concepts.

However, Stitch’s feature set is intentionally streamlined. It doesn’t aim to cover the full spectrum of design and development workflows—such as backend logic or advanced animation—but instead focuses on making the first steps of UI creation radically faster and more accessible.

Stitch delivers a practical shortcut from idea to interface, letting teams skip repetitive setup work and focus on refining the user experience. As Google continues to expand its features, Stitch is likely to become a go-to tool for anyone looking to move from concept to clickable UI with minimal friction.