Google 演示在 Gemini Canvas 中用草图生成应用原型

Google published a Gemini walkthrough on November 5, 2026, showing how a photograph of a paper sketch can become an interactive app in Canvas. The demonstration covers uploading the image, generating code, testing the result and refining it through conversation.

The starting point can be an app mock-up, habit tracker or game board. After the image is uploaded, the user selects Canvas and asks Gemini to code the app. In the demonstration, Gemini reads the visual and writes the code. The resulting tool can be tested in a preview window, followed by real-time requests to change colors, fonts or features.

Canvas itself predates the walkthrough. Google introduced it on March 18, 2025, as an interactive space for documents and code, with support for generating and previewing HTML/React and other web app prototypes. The sketch example uses that established prototyping environment with an uploaded image as the starting point.

Editing and Testing the Generated App

Google’s desktop Canvas documentation supplies the controls around the short demonstration. Canvas requires a signed-in Gemini Apps user, and an image or files can accompany the initial prompt. Updates can be requested conversationally or made by editing the code directly, with changes autosaved.

On desktop, Code view opens the generated code for direct editing. Show console exposes errors and logs from the preview, while a recent-changes view provides a way to inspect code revisions. Select & ask lets the user choose a section and describe the desired revision, allowing a targeted request alongside broader chat instructions.

The mobile instructions also provide Preview and Show console controls. Android users can add a home-screen shortcut for a Canvas app. Signed-in users aged 18 or over can select Add Gemini features to incorporate Gemini-powered text or image generation, then continue requesting changes in chat.

Availability and Public Sharing

Sharing extends the workflow beyond the creator’s own preview. Canvas apps can be shared through a public link, with support for sharing app data between users and saving data across sessions and devices. Anyone with the public link can view and edit data associated with the app, making the link an access decision as well as a distribution method.

Shared Canvas links open only on the web, not in the Gemini mobile app. Recipients can copy a shared Canvas and continue editing it themselves, although that function is unavailable to users under 18. If the recipient’s Keep Activity setting is enabled, opened shared Canvas content is stored in Gemini Apps Activity.

The walkthrough shortens the route to a first version. The work that follows sits in the preview, code history, console and sharing controls: checking generated behavior and deciding who receives a link that exposes editable app data. A sketch can initiate the prototype, but wider use depends on those less visible inspection and access decisions.