Google is testing a new browser concept called Disco , an experimental offering that combines classic web browsing with generative artificial intelligence. It's not yet intended as a direct replacement for Chrome, but rather as a testing ground where the company can experiment with ideas that could shape the future of how we use the internet.
Instead of simply adding small AI features to Chrome, the company has decided to explore a different browser, designed from the ground up to work with AI . Disco relies on Gemini 3, Google's most advanced language model, to interpret your web activity, understand your goals, and transform individual tabs into complete tools without requiring any coding.
What is Google Disco and how does it differ from Chrome?
Disco was born within Google Labs as an experimental project focused on intelligent navigation. Technically, it's built on Chromium , the same open-source codebase that powers Chrome, similar to offerings like Opera Neon , but its philosophy is radically different: it doesn't just display pages, it attempts to interact with them.
The goal is for the browser to move beyond being a passive content viewer and become an assistant that understands and anticipates your needs . It's no longer just about opening links and reading; it's about Disco processing what you have open, connecting the information, and generating practical results, such as comparisons, summaries, or interactive dashboards.
In this approach, each tab becomes an active element that AI analyzes in real time . The information is not isolated on each website, but is combined to fulfill the user's request: organizing a trip, comparing products, summarizing lengthy documents, or designing custom tools.
For Google, Disco serves as a parallel testing ground for Chrome . If the ideas take hold and prove useful, many of these AI features will likely end up in the browser already used by billions of people worldwide.
A browser where every tab is “alive”
The big news in Disco is that tabs are no longer just static pages . Every site you open is analyzed by Gemini 3, which interprets the content, structures it, and connects it with the other open tabs when it makes sense to do so.
This architecture allows the browser to, for example, relate information from multiple websites simultaneously . Instead of the user jumping from one tab to another, Disco can work with all of them together to offer a compact and actionable view.
Imagine you're browsing through different stores and hardware reviews: graphics cards, RAM, laptops, or PC components. The idea is that you can ask Disco something as simple as: "compare these options and tell me which one offers the best value for money and performance ." The AI cross-references the data from all the tabs, extracts specifications, prices, and reviews, and generates a unified view.
Something similar happens with other everyday tasks, such as trip planning. If you have tabs open with flights, hotels, and travel guides, Disco can combine all that information into an organized itinerary with dates, maps, and points of interest, without you having to piece it all together by hand.
In practice, this means that navigation is shifting from "I read all the websites" to a more "I set the goal and AI takes care of weaving the content together" approach . The user can still access the original sources whenever they want, but is no longer obligated to review everything step by step.
GenTabs: turning tabs into interactive mini-applications
The heart of Disco is called GenTabs . This feature is responsible for transforming several loose tabs into something more sophisticated: small, interactive web applications created on the fly , without the need to write code.
The workflow is relatively simple. First, you navigate as usual: you open the pages with the information you're interested in, whether it's articles, stores, maps, or technical documentation. Then, you describe what you need in natural language : for example, "organize a three-day itinerary using these websites," "give me an understandable summary of this explanation of entropy," or "create a comparison tool for these laptops."
This is where Gemini 3 comes in. It analyzes the relevant tabs and generates a GenTab—a new tab tool that presents the results in a more visual and manageable way. Instead of being limited to plain text, it can structure tables, calendars, maps, interactive lists, or basic controls for adjusting parameters.
Google explains that you don't need any programming knowledge or to touch a single line of code . You simply refine the tool with new instructions in natural language: ask it to add fields, change the order of the information, filter by specific criteria, or simplify overly technical parts.
Each GenTab maintains direct links to the original web sources . This allows you to verify the data, expand your knowledge, or review the context from which the AI drew its conclusions—important if you plan to use the results for a report, academic paper, or purchasing decisions.
Practical examples: from scientific explanation to organized trip
Among the examples that Google has shown in its internal demonstrations and official communications are quite varied cases, which help to get an idea of how far this type of AI browser could go.
In the educational field, Disco can take a dense webpage about a concept like entropy and generate a visual application that explains it more intuitively . This mini-app can include interactive representations, step-by-step explanations, and summaries that can then be reused, for example, for studying or writing a paper.
Another common use case is trip planning . Using flight, hotel, and travel guide websites, Disco can create a kind of control panel that centralizes everything: a map of the area with key points of interest, integration with the user's calendar, an itinerary by days and time blocks, and references to selected locations. The idea is to avoid having to copy information from one tab to another.
More domestic scenarios have also been demonstrated, such as a meal planner or a garden organizer . The browser can group recipes, product information, planting calendars, and maintenance tips into a single interactive interface, instead of leaving the user to manage with multiple bookmarks and loose notes.
In all these cases, the logic is the same: use the web as raw material and AI as assembler , so that the final result is a useful and specific tool for the task that needs to be solved.
Proactive navigation and collaboration with AI
With Disco, Google is focusing on a much more proactive and less manual browsing experience . Instead of opening dozens of tabs and delving into each one, the idea is that you open what you need and then ask the AI to generate the result you're looking for.
This change transforms the browser into something more like a digital work partner that understands context, objectives, and relationships between data. It doesn't just answer isolated questions like a traditional search engine; it takes into account what you already have open, how you're using it, and what task you're trying to complete.
For many advanced users, this could translate into faster workflows : journalists putting together dossiers from multiple sources, professionals preparing product comparisons, or citizens who want to plan procedures, trips, or personal projects without wasting time jumping between tabs.
Google insists that each generated result maintains traceability back to the web, so the link to the original pages is not broken . The browser becomes an organizer and orchestrator, but it does not completely replace direct access to the sources.
In any case, the company acknowledges that this technology is still in its infancy , with clear limitations: the results won't always be perfect, errors or inaccurate interpretations may occur, and the quality depends largely on the reference websites chosen by the user. Therefore, they recommend being cautious with the data entered, especially if it is sensitive.
Availability, access and situation in Spain and Europe
Regarding access, Disco is currently in a restricted experimental phase within Google Labs . The initial test is being conducted with a limited number of users, primarily in the United States, and for now, the available version focuses on macOS devices.
Those who want to try the browser must sign up for a specific waiting list through the Google Labs website for Disco. Once accepted, they can download the client and begin submitting feedback on their experience, something the company considers crucial for deciding which features to keep, which to modify, and how to prioritize its release on other platforms.
For users in Spain and the rest of Europe, this means that direct access is, for now, very limited . Google will likely adopt a phased approach, expanding to more countries and operating systems as it delves deeper into issues such as regulation, data privacy , and adaptation to European legal frameworks.
Although there are no official dates for a general release, it's almost a given within the tech community that some of the features tested in Disco could eventually be incorporated into Chrome . If that happens, European users could benefit from these features without needing to install an additional browser, provided that EU data protection and transparency regulations are met.
Google has also indicated that Disco is designed as a platform that will add new AI features over time. GenTabs is just the first piece; the company suggests that other experimental capabilities will be rolled out to testers already in the early access program.
A laboratory for the future of AI-powered navigation
Beyond its current limitations, Disco represents an alternative to the general trend of AI-powered browsers . While other companies are opting to integrate assistants directly into their classic browsers—such as Microsoft Edge's copilot mode or offerings like ChatGPT Atlas —Google has chosen to separate the experiment into a distinct product.
This separation allows them to test riskier ideas without abruptly altering the experience of Chrome's 3.800 billion users . If something doesn't work as expected, it can be refined on Disco; if it's successful, they can then assess how to integrate it into the main browser. At the same time, it allows them to compare approaches with external projects like Comet.
At the same time, the company emphasizes that Disco is not yet ready for unsupervised, critical or professional use . The company itself describes it as being in an early stage, with imperfect performance and significant room for improvement, and recommends caution with the data entered, especially if it is sensitive.
For European users, businesses, and developers interested in this evolution, Disco can serve as a clear indicator of where the industry is heading : fewer clicks, more context; fewer mechanical tasks, more AI-based automation; and browsers that understand, at least in part, what you want to achieve and not just what you see on the screen.
Disco is shaping up to be an ambitious experiment with which Google is exploring how to combine generative AI and web browsing . If it manages to overcome its limitations and find the right balance between usability, transparency, and respect for privacy, many of the features currently being tested with a small group of users could eventually become part of the daily lives of those browsing from Spain, Europe, and the rest of the world.
