On December 11, Google made an announcement about the launch of its experimental AI tool, 'Disco', within Google Labs. This innovative tool harnesses the power of the Gemini model to scan the content of users' currently active browser tabs and seamlessly merge it to create interactive web applications.

Users have the ability to craft 'GenTabs' applications by simply using natural language prompts. Additionally, Disco takes a proactive approach by offering application-building suggestions tailored to the user's current browsing task. For instance, when a user is delving into a specific topic, the system can recommend the creation of visualization tools to display pertinent knowledge through charts or structured views. When browsing through multiple recipes, GenTabs can automatically generate interactive meal plans. And when searching for travel information, it can consolidate destination details, itineraries, and price data scattered across various tabs into a practical itinerary planning application.

GenTabs will tap into the Gemini 3 model to dynamically generate complete application interfaces and logic as needed. This generation is based on the content of the browser tabs and the user's conversation history. Moreover, it allows for iterative modifications and fine-tuning of the generated applications through natural language interactions. At the same time, the generated content maintains links to the original information sources, enabling users to trace the origins of the data.

Currently, Disco is exclusively accessible to a select group of early testers within Google Labs, with the initial rollout occurring first on the macOS platform. Users interested in trying it out must join the waitlist via the relevant Google Labs page.