Google Launches Stitch: AI-Powered App Design Tool

At Google I/O 2025, Google unveiled Stitch, an AI-powered tool designed to simplify web and mobile app front-end development. Stitch generates UI elements and code, streamlining the design process.

Users can create app UIs with simple text prompts or even images. Stitch then provides HTML and CSS markup for the generated designs. The tool leverages the power of Google's Gemini 2.5 Pro and Gemini 2.5 Flash AI models for code and interface ideation.

Stitch Interface
Stitch lets users choose between Gemini 2.5 Flash and Gemini 2.5 Pro models. Image Credits: Jagmeet Singh / TechCrunch

AI-Powered Coding Gains Momentum

Stitch enters a growing market of "vibe coding" tools that utilize AI models for code generation. Several tech startups are exploring this space, including Cursor, Cognition, and Windsurf. OpenAI recently launched Codex, and Microsoft updated GitHub Copilot, further highlighting the trend.

While Stitch offers a focused approach compared to some competitors, it provides valuable customization options. These include direct export to Figma and code exposure for refinement in an IDE. Users can also fine-tune generated design elements.

Streamlining Initial Design Iterations

In a demonstration, Google product manager Kathy Korevec showcased Stitch creating a mobile UI for a book app and a web dashboard for beekeeping.

Stitch helps you complete your initial iteration quickly and easily continue from there. We aim to make design thinking and software building more approachable for everyone.

Google plans to add a feature allowing users to modify designs using annotated screenshots. While powerful, Stitch is not intended to replace full-fledged design platforms like Figma or Adobe XD.

Stitch Prompts
Stitch is a focused tool for initial design iterations. Image Credits: Jagmeet Singh / TechCrunch

Jules: AI-Powered Bug Fixing Enters Public Beta

Google also expanded access to Jules, its AI agent for bug fixing. Now in public beta, Jules assists developers in understanding complex code, creating GitHub pull requests, and managing programming tasks. A demonstration showed Jules upgrading a website from Node.js 16 to Node.js 22, including verifying functionality after the upgrade.

Jules currently uses Gemini 2.5 Pro, with future support for other models planned.