Google LLC is adding features to its artificial intelligence Gemini chatbot today that will allow it to collaborate with users in a new interactive canvas for documents and code.

“Canvas simplifies the entire coding process, allowing you to focus on creating, editing and sharing your code and design in one place, without the hassle of switching between multiple applications,” said Dave Citron, senior director of product management for Gemini apps at Google.

When activating Canvas, once users prompt Gemini to generate code by asking it to begin writing an app or document, it will display changes in a sidebar in real time.

For document editing, this means users can quickly and easily adjust their text directly in the document without needing to “talk” to the chatbot conversationally each time they want a change. This style of editing can be tedious and slow, making iterating through a document plodding when asking, “Please change the sentence of the second paragraph.” Instead with Canvas, a user can adjust the sentence directly.

Sentences and paragraphs can also be highlighted, allowing users to trigger Gemini with a prompt to modify them. According to Google, this will permit users to collaborate deeply with more nuanced editing decisions with text – including both research and fine-tuning their documents.

Once a user is done editing a document in Canvas they can easily export it into Google Docs where they can continue collaborating on their text with their coworkers.

For code, developers can ask Gemini to write code that will also appear in the sidebar, where they can edit alongside Gemini. For web apps, it can visualize changes in HTML, CSS and React JavaScript, allowing software engineers to collaborate with the chatbot.

A solar system simulation JavaScript web app made using Gemini with only three prompts.

As the chatbot changes the web app, the user can interact with the app in the canvas including click buttons, watch animations and play around with the user interface. For example, it could be used to make a simple simulation of the solar system and include statistics for each of the planets.

Web apps generated by the Gemini app can run from relatively simple to extremely complex, depending on the amount of time and fine-tuning the developer wants to spend talking to the AI. Advanced developers and beginners with little or no experience can create complicated web applications in only a few minutes of conversation with the chatbot.

During a presentation led by Google, Citron said that this is just another step in the company’s work toward making the Gemini app more “agentic.”

“In general, this Canvas ability is yet another part of making the Gemini app increasingly agentic,” Citron explained. “And by agentic, I mean this idea that AI is working on your behalf to get things done and not just simple things like turning on the lights, but more and more incredible things that I didn’t even think was possible with AI until a couple of years ago.”

Canvas is available today to all users of Gemini globally.

Bringing audio overviews to Gemini

Audio Overview transforms documents, text and extensive notes or reports from Gemini’s Deep Research tool into an engaging, podcast-like audio discussion featuring two AI-voiced speakers.

In this format, the speakers engage in a witty back-and-forth, reminiscent of a talk show, as they summarize the topic, connect key points and share light-hearted banter.

This capability was introduced by Google in 2024 as part of NotebookLM, the company’s AI-powered note-taking and research assistant. It functions similarly to the research assistant, enabling users to listen to the audio while their hands and eyes are occupied with other tasks, such as driving or doing chores, allowing them to stay informed with AI-generated information.

Google announced that the Audio Overview feature is now available globally to Gemini and Gemini Advanced subscribers, initially in English, with plans to roll out additional languages in the coming weeks.

Images: Google, SiliconANGLE/Gemini