Google’s latest Stitch update transforms it into a powerful AI-driven design tool, featuring a native canvas, customizable Gemini agents, and seamless integration with platforms like Figma, React, and AI Studio. Key features include extracting design standards from websites, a design.md system toolkit, voice-driven editing, and instant prototyping, making Stitch accessible for non-designers and free to use.
Google Labs has released a major update to Stitch, its generative AI design tool, transforming it from a simple screenshot utility into a robust competitor to tools like Figma. The latest version introduces an agentic system that leverages both Gemini text and image models, allowing users to prompt the tool to collaboratively build designs. Stitch now features a native design canvas, similar to AI Studio, where users can generate and edit designs through prompts, and it introduces a new design agent that can be customized with different Gemini models (such as 3 Flash or Pro).
A standout feature of this update is the introduction of a design.md file, which acts as a comprehensive design system toolkit. This file can be edited graphically within Stitch or exported for use in coding editors or other platforms, making it easy to apply consistent brand standards across multiple projects. Additionally, Stitch can now extract design standards directly from a website URL, automatically pulling colors, fonts, and styles from existing sites to inform new designs.
Stitch also enhances workflow integration by allowing users to export their designs directly to AI Studio, where they can be turned into functional apps with authentication, databases, and more. The tool supports various export options, including Figma and React, and can generate instant prototypes that users can preview across different device formats. The system also supports voice-driven design through Gemini Live, enabling users to make design changes simply by speaking to the interface.
The design process within Stitch is highly interactive and flexible. Users can generate multiple design variations, preview and wireframe pages, and make AI-driven edits to specific elements. The tool uses advanced image generation models to create realistic placeholder images and allows users to upload their own assets. Once satisfied, users can export their work, generate a project brief or product requirements document, and access or modify the underlying design system for further customization.
Overall, this update positions Stitch as a powerful, accessible design tool for users who may not be professional designers but have clear ideas and references for their projects. By integrating closely with Google’s latest AI models and offering seamless export and prototyping features, Stitch is evolving into a versatile platform for app and web design. The video’s creator notes the potential for future expansion into other design areas and encourages viewers to try Stitch, highlighting that it remains free to use.