Stitch, launched by Google at Google I/O 2025, generates user interface designs from a text prompt, sketch, or reference screenshot, using Gemini's multimodal models to interpret what's being asked for and build a responsive interface in an AI-native design canvas rather than a traditional static mockup tool.
A 2026 update pushed it further into live, collaborative design: Stitch Agent can now stream work onto the canvas in real time while a user steers the direction, import existing design or code files, and generate motion-focused interfaces. Output exports directly into Figma for further design work, as HTML and CSS code for developers, or to Google Antigravity, also covered in this directory, extending Google's coding tools.
Stitch is free as part of Google Labs, with generation limits depending on the model mode: up to 350 monthly generations on the faster Gemini 2.5 Flash mode, or 50 on the more detailed Gemini 2.5 Pro mode. For a designer or product team already working with Gemini or Google's broader AI tooling, Stitch's direct export into Figma and Google's own coding tools connects design and implementation more directly than a standalone design tool would.
