IA
Google shows how to turn a sketch into an application with Gemini Canvas
Google offers a simple stream in Gemini: import the photo of a sketch — application model, tracking tool or game, for example — then open Canvas and ask the AI to code it. Gemini interprets the visual, generates the code and displays an overview in which the user can test the result, change colors, typography or functions, and continue to iterate. The official documentation confirms that Canvas accepts adding images and allows to edit the code directly, examine errors and share an application. A Gemini account is required; Google does not present this prototype as a product ready to be put into production.