Google Gemini: Two new very promising functions with Canvas and Audio Overview
The current period is definitely very intense in terms of all -out advertisements concerning language models. After GPT 4.5, Mistral Small 3.1 and several models at Alibaba And Baiduit's Google's turn to go with its new features. He had started a few days ago with Gemma 3on the open source front. On the side of its proprietary models, the Gemini 2.0 being very recent, the web giant presented them rather new functions for its gemini: Canvas, an interactive space dedicated to collaboration on documents and code, and Audio Overview, which converts files into audio discussions in Podcast.
Canvas: a collaborative environment for writing and code
Let's start with Canvas, which presents itself as an interactive space integrated directly into Gemini, and which is not without removing the “canvas” proposed in Chatgpt. It allows you to write, modify and structure documents or code without leaving the assistant's interface. Its strength lies in real -time collaboration with Gemini, which can generate a first version and then offer adjustments to style, length or formatting. It is a kind of textual sandbox in which it is therefore possible to exchange more easily and more fluid with AI.
Access is simply via Gemini's entry bar by selecting “Canvas”. The user can then refine its content by requesting specific modifications, such as making a passage more synthetic or adapting the tone. Once the document is finalized, it is possible to export it to Google Docs in one click. Google suggested that the possibility of exporting other types of documents would be added later.
Canvas is also aimed at programmers by simplifying the transition from an idea to a functional prototype. It supports several languages and mainly offers a real -time overview of the HTML/React code, especially for web interfaces. A user can for example create a subscription form, then ask Gemini to generate the HTML code and instantly visualize it in stride. Any modification – Adding a field or button – is immediately applied in the overview. Such a mode of operation avoids having to juggle between several applications and should significantly accelerate development. Another use case that appears obvious to us is that of programming students, who will be able to benefit from a more visual and interactive learning.
Overview audio: transform documents into podcasts
On the sidelines of this Canvas, Google also incorporates Gemini an improved version of Overview Audio, a feature that generates audio discussions from text files. It is therefore the same technology that had amazed us in Notebooklm. She therefore invites herself in Gemini to convert reports, slides and articles into conversations analyzed by virtual presenters.
The user only has to download a document: with one click, Gemini produces an oral summary in the form of an exchange between two artificial votes which comment and connect the main information. A format designed to assimilate complex content fluidly.
Practical to listen to a document by performing other tasks, Audio Overview is accessible from the web and the Gemini mobile application. It also saves and share the discussions generated.
Availability and careful languages
These new features are accessible today for the Gemini and Gemini Advanced subscribers. Canvas is available in all languages supported by the application, while Audio Overview is currently limited to English, with other languages planned in the long term.
To use Canvas, simply select it in the input bar. As for Audio Overview, it activates after importing a file, via a suggestion displayed above the entry area.
