Google Disco (labs.google/disco), developed under Google Labs in collaboration with the Chrome team, is an experimental AI-native web browser designed to shift web navigation from passive tab consumption to active task execution. Powered by Gemini foundation models, Disco’s core feature—GenTabs—analyzes a user’s open tabs, browsing intent, and chat prompts to dynamically synthesize interactive, task-specific mini web applications (such as live travel planners, dynamic product comparison dashboards, and interactive study suites) directly inside the browsing window with transparent source attribution.

Traditional web browsers function primarily as document viewers where users juggle dozens of disconnected tabs, manually copy-pasting information across notes, spreadsheets, and external apps. Project Disco inverts this interaction model by turning open web pages into raw components for Generative User Interfaces (GenUIs). Instead of merely summarizing web pages in an isolated sidebar chat, Disco orchestrates multi-tab context into unified, interactive workspace tools.

Core Architecture & Technical Profile at a Glance

Feature / Dimension	Google Disco Specification	Primary User & Workflow Advantage
Core AI Engine	Gemini Foundation Models (Multimodal Context)	Deep semantic understanding of open tabs, DOM structures, and multi-page intent.
Pioneering Feature	GenTabs (Generative Task-Specific Web Apps)	Synthesizes interactive tools from scattered tabs via natural language commands.
Browser Foundation	Chromium-Based Architecture	Full compatibility with standard modern web standards, extensions, and rendering.
Organizational Unit	Task-Based “Projects” Workspace	Groups open tabs, conversation history, and generated GenTabs per workflow.
Source Attribution	Interactive Bidirectional Deep-Linking	Every UI element, table row, and chart maintains active links to underlying source pages.
Target Availability	Google Labs Experimental Waitlist (macOS Preview)	Limited beta release for research, user feedback, and workflow experimentation.

Key Capabilities and Practical Workflows

• Dynamic Mini-App Generation (GenTabs): When researching a destination, shopping across multiple stores, or gathering academic papers, users can prompt Disco to organize the information. GenTabs builds interactive sorting filters, budget calculators, interactive maps, or comparison matrices pulling live data from open tabs.

• Living, Dynamic Web Workspaces: Unlike static text summaries, GenTabs remain connected to the browser state. Opening an additional flight option, hotel link, or review page automatically prompts the GenTab to update its interactive itinerary or comparison table with the newly discovered data.

• Proactive Contextual Suggestions: Disco identifies workflow gaps based on user actions—suggesting a budget tracker during vacation planning, a citation formatter during literature reviews, or a feature matrix during software evaluations.

• Web-Grounding and Source Preservation: Rather than keeping users locked in an abstract chat window, Disco encourages deep engagement with primary web content, ensuring all synthesized interface components link directly back to original publisher URLs for transparency and verification.