Google announced that the Canvas function previously used in Google Labs experiments is officially open to all US users through Gemini’s AI Mode (AI mode), and currently supports English use. Designed to help users better organize information, plan projects, and conduct in-depth research, Canvas now extends to the ability to compose documents and create custom tools within the search interface.
According to Google's introduction in the blog, users can use Canvas in AI mode to draft various manuscripts, or directly generate tools suitable for specific tasks in Google search. In the past, Google has recommended using Canvas for creating study guides, such as uploading class notes and other materials, and having the system automatically generate structured content. Canvas can also convert research reports into web pages, quizzes, or audio overviews, which has some functional overlap with Google's research tool Notebook LM.
In the creation and development scenario, users only need to describe their ideas to Canvas in natural language, and they can get the corresponding code and quickly turn it into a shareable application or mini-game. This feature is also applicable to literary creation and various creative projects. Users can use Canvas to polish their drafts, adjust their structure, or get suggestions for improvement.
Currently, Canvas has been integrated into the Gemini system. Users who subscribe to Google AI Pro and Google AI Ultra can use the latest Gemini 3 model and obtain a context window of up to 1 million tokens to handle more complex and information-intensive projects. As Canvas is released to all users in the United States through AI Mode, many users who have not actively tried Gemini before will also be exposed to this feature in the search experience for the first time. Google hopes to rely on the huge coverage of search products to present its capabilities to more users in this AI competition.
In terms of the specific usage process, the user clicks the “+” button in the tool menu in AI mode, selects the newly appeared Canvas option, and then describes the task he wants to complete or the content he wants to create. This will open a Canvas panel on the side of the interface, where users can aggregate information from web pages and Google Knowledge Graph to organize research materials or structure projects. If you are building a prototype or app, users can test functionality in the same interface, switch between views of the underlying code, and continuously fine-tune the app's behavior through conversations with Gemini.
In terms of positioning, Canvas also faces direct benchmarking from competitors such as OpenAI and Anthropic. For example, the Canvas function in ChatGPT is automatically triggered based on user questions, while Claude from Google and Anthropic places more emphasis on users actively selecting or evoking relevant tools. These products all support writing assistance, project planning, and transformation “from idea to finished product,” but they differ in interaction paths and triggering methods.
Currently, Canvas in AI Mode is only available in English for users in the United States, which can be used to create plans, projects, applications and other content. Google has not revealed a further timetable as to whether it will be expanded to more languages and markets in the future.