Gemini can now turn a photo of your sketch into a working app

Google says its Gemini app can now turn a photographed hand drawing into a working interactive app. In a post published November 5, the company describes a three-step flow built on Gemini’s Canvas tool.

Three steps from paper to a running prototype

According to Google, the process starts with a photo of a sketch uploaded to Gemini. The post gives a few examples of what the drawing could show: an app mock-up, a habit tracker or a game board. The user then selects the Canvas tool and asks Gemini to build the app from the image. Google’s suggested prompt is short: “Can you code this app?”

Gemini reads the drawing and writes the code, and the result appears in a preview window. From there, the user can ask for changes to colors, fonts or features. Google says Gemini can make those adjustments “in real time.”

The post is brief and works as a how-to rather than a technical release. It does not name a model, describe how well Gemini handles messy handwriting or complex layouts, or show benchmark results.

What the announcement leaves unsaid

Google’s page does not say who gets the feature, on which platforms, in which regions or languages, or whether any paid plan is required. It also gives no usage limits. A separate Google post from October 29, about generating and editing files inside Gemini, said those capabilities are available to all users in every language where the Gemini app runs, with some formats limited to the web. That statement does not cover this feature, and readers should not assume the same terms apply.

The feature also is not a new product. Canvas, the workspace where Gemini writes and previews code, has been part of the app since Google I/O 2025, according to coverage by Scalevise and others. What the new post adds is an image as the starting point instead of a text description.

A shortcut for prototypes, with limits for real products

For a non-programmer, the practical change is the input. Describing a layout in words is often harder than drawing it, and a photo of a whiteboard or notebook page skips that translation step. That should suit quick experiments such as a personal tracker or a simple game, which are the kinds of examples Google picked.

Outside reporting suggests where the ceiling is. Tutorials summarized in recent search coverage say Canvas builds client-side apps, so databases, sign-in and server logic have to be added separately. Those same sources say apps that pull in Workspace data cannot be previewed. Scalevise draws a line between Canvas, which it describes as suited to rapid visual prototyping inside Gemini, and Build mode in Google AI Studio, which it says generates a code-backed, full-stack web app and is the better route for something deployable. These are third-party descriptions of Canvas generally, not statements from Google about the sketch feature, and the limits may have changed since they were written.

Google’s claim that the app reflects the drawing is also unverified. The post shows no side-by-side comparisons of sketches and outputs, so how faithfully Gemini reproduces a hand-drawn design is untested in public material.

Part of a steady run of Gemini app additions

The sketch feature arrives days after Google said users can generate, preview and edit documents, PDFs, presentations and spreadsheets directly in Gemini. Together, the posts show the company filling the Gemini app with in-chat workspaces, so that a conversation can end in a finished file or a running app without leaving the window.

Google’s announcement gives no roadmap for the feature. Whether sketch-to-app expands to more input types, connects to AI Studio for deployment, or ships with usage caps is not addressed. Until Google publishes availability details, the safest reading is that the flow is described for the Gemini app with Canvas, and that anyone interested should test it in their own account to see whether it appears.