Google has introduced a new experimental design tool called Stitch, which uses generative AI to turn simple text prompts and images into fully designed user interfaces. Unveiled during the company’s annual developer conference, I/O 2025, Stitch is being positioned as a major step forward in streamlining app prototyping and front-end development workflows.
Designed by Google Labs, Stitch is powered by Gemini 1.5 Pro and Gemini Flash—Google’s latest large language models—enabling the tool to rapidly interpret user intent and generate visually coherent layouts, front-end components, and even production-ready HTML/CSS code.
A New Era of Design, Powered by Language
Stitch allows users to describe the look and feel of an app screen in natural language. For example, a prompt like “Create a travel booking app homepage with a minimal theme and ocean blue tones” results in an interactive mockup and code that can be exported directly.
The tool also supports image-based input: users can upload sketches or screenshots, and Stitch will convert these into polished UI drafts. This image-to-design capability is especially helpful for early-stage ideation, removing the need to rely solely on traditional design software or manual layout work.
According to Google, the AI can understand abstract instructions such as “sleek, professional dashboard for fintech” or “friendly UI for a kids’ learning app,” and generate design elements aligned with the desired tone and audience.
Seamless Export and Figma Compatibility
One of Stitch’s standout features is its integration with other design and development platforms. Users can export their AI-generated layouts directly to Figma for further editing or refinement. Alternatively, front-end developers can export the layout in usable HTML and CSS, bridging the often-fragmented workflow between design and development.
“Stitch was built to collapse the time between idea and execution,” said Eli Collins, VP of Product at Google DeepMind, during the I/O 2025 keynote. “Whether you’re a startup founder sketching your first interface or a developer validating user flows, Stitch accelerates the creative loop.”
Not Without Limits
While powerful, Stitch is not yet a full design suite replacement. The tool currently excels at generating single-screen UI mockups but lacks robust support for complex, multi-screen user journeys or advanced interactive prototypes. Some designers have also noted its limited options for customizing fonts, spacing, and accessibility elements, making it more suitable for ideation than production design in its current state.
Additionally, design fidelity—the precise alignment with specific branding or style guides—may require manual adjustments after initial generation.
Open Access via Google Labs
Stitch is now available for public experimentation via stitch.withgoogle.com, under the Google Labs umbrella. As with other Labs projects, Stitch is being rolled out gradually, and users are encouraged to share feedback to help shape its development.
In the broader context, Stitch represents Google’s expanding push into AI-assisted creativity tools. It joins other initiatives such as MusicFX for sound design and VideoFX for AI-generated video editing, illustrating Google’s strategic commitment to AI-enhanced productivity across disciplines.
Conclusion
Stitch is a bold attempt to reimagine how digital interfaces are conceived and constructed. Though still experimental, it shows significant promise in reducing the time and effort needed to move from concept to clickable interface. For developers, designers, and product teams, Stitch could become an indispensable tool in the rapid prototyping and UI ideation toolkit.
As generative AI continues to evolve, tools like Stitch highlight a growing trend: designing software may one day require more imagination than instruction.
Sources: