On Thursday, Google unveiled a new web browser AI experiment named Disco, powered by Gemini. This tool transforms your open browser tabs into personalized applications. Through Disco, users can generate what Google refers to as “GenTabs”—a feature that actively recommends interactive web apps to assist with tasks related to your current browsing. It also enables you to construct your own apps using simple written prompts.

For example, while studying a topic, GenTabs might propose building a web app to visualize the information, aiding in grasping key concepts more effectively. In everyday situations, you could use it to devise a meal plan from various online recipes or organize a trip while researching travel destinations.

While similar tasks can already be performed with certain AI chatbots, GenTabs creates these tailored experiences instantly using Gemini 3, drawing from your browser activity and Gemini chat history. Once an app is created, you can further adjust it using natural language instructions. Google also mentions that any generative content within GenTabs will reference the original sources.

Following industry trends, Google has been exploring ways to integrate AI more deeply into web browsing. Rather than developing a separate AI browser like Perplexity’s Comet or ChatGPT Atlas, Google embedded its Gemini assistant directly into Chrome, where it can optionally answer questions about the webpage you’re viewing.

With GenTabs, the emphasis expands beyond the current page to encompass your broader browsing activity across multiple tabs—whether for research, learning, or other purposes. Initially, this feature will be accessible only to a limited group of testers via Google Labs, who will provide feedback on their experience.

Google indicates that promising concepts developed through Disco could eventually be incorporated into other major Google products. It also notes that GenTabs is just the “first feature” being tested, with more Disco capabilities expected in the future. To try Disco, users can join a waitlist to download the app, beginning with macOS availability.