Google launches Disco, GenTabs turn tabs into web apps

Google Labs launched Disco on Dec. 11, 2025, an experimental generative-AI web app that uses the Gemini 3 model to transform a user’s open browser tabs and chat history into interactive GenTabs web apps that help complete complex tasks without coding, available initially to limited testers via labs.google to speed discovery, planning, and collaboration.

Google has launched Disco, an experimental web browser

What is Disco and who should care? Google Labs positioned Disco as an experiment that sits at the intersection of search, browsing, and generative AI, aimed at professionals who juggle many sources and need structured tools fast. The company rolled the experiment out to an early group of testers on Dec. 11, 2025, and described the project as a way to let AI synthesize session context into lightweight apps called GenTabs.

Why this matters: Disco reframes the browser session as raw material. Instead of bookmarking and manual note-taking, Disco analyzes open tabs and recent Gemini chat interactions to propose ready-made tools—planners, comparison tables, annotated bookmarks—that would normally take hours to assemble. Early internal demos suggested time savings of 30–60% on tasks like trip planning or research summarization.

Examples and data-driven cases make the change tangible:

• Trip planning: Disco generated a planner combining flight options, calendar slots, and a map in about 90 seconds versus 20–40 minutes manually.
• Research synthesis: A market analyst converted five product pages and two PDFs into a comparison dashboard with key metrics highlighted.
• Learning module: A student turned a set of tutorial tabs into an interactive study guide with flashcards and example prompts.

Actionable insights for teams experimenting with Disco:

• Start with a focused goal: limit tabs to one project to improve GenTab relevance.
• Iterate prompts: refine natural-language instructions to the GenTab for tailored outputs.
• Export early: use Disco outputs as templates to standardize workflows across teams.

Table: Example GenTabs and estimated time savings

How GenTabs works in Disco

GenTabs are Disco’s central feature: small, interactive web apps generated from a user’s browsing session and Gemini chat history. They convert context into widgets, tables, timelines, and even simple 3D previews that users can refine via natural language prompts.

How GenTabs work

GenTabs follow a three-step flow. First, Disco analyzes open tabs and recent Gemini 3 chat context to extract entities, intents, and resources. Second, a generation step constructs a focused app skeleton—fields, visualizations, and suggested actions—matched to the detected goal. Third, users iterate by typing commands or accepting smart suggestions to refine layout and content. The underlying Gemini 3 model handles extraction and synthesis, while Disco frames the UI as a manipulable app rather than a static response.

Concrete examples clarify the mechanics: when a user opens multiple hotel pages and a calendar, Disco can propose a GenTab that lists options, overlays available dates, and flags potential conflicts. When a product manager gathers specs across vendors, Disco generates a comparison grid with sortable columns and recommended next steps.

Availability & controls

Disco launched to a limited cohort through Google Labs, accessible via labs.google. Early access is gated—testers sign up and receive invites—so adoption will scale progressively. Privacy and control surfaced repeatedly: Google added session-level toggles that let testers choose which tabs or chat segments Disco may ingest. There are manual override controls and export options to copy GenTab data to Google Drive or download CSVs for team workflows.

• Control tips: keep project tabs grouped to reduce noise.
• Privacy tip: use session exclusion for sensitive tabs.
• Collaboration tip: export GenTabs to share structures with stakeholders.

Google Labs & experiment history

Google Labs is the company’s sandbox for playful, community-driven AI features. Disco follows earlier Labs prototypes that blended models with search, and it sits alongside experiments that explore how AI can reshape routine tasks. Labs emphasizes iteration: Disco’s limited release mirrors past launches where Google gathered tester feedback before wider rollouts.

Comparison table: Disco vs. mainstream browsers and AI tools

"Disco shows how browsing can produce tools, not just results," said an early tester, describing faster handoffs between research and execution.

What's Next / Future Implications

Disco’s debut suggests a larger shift: browsers may evolve from passive viewers into active collaborators. For professionals, the key implications include faster prototyping, less manual aggregation, and new patterns for team handoffs. Google has framed Disco as an experiment, so the timeline depends on tester feedback and privacy reviews before any broader release.

AI + browsing trend

How this trend unfolds will shape tools for developers, analysts, and strategists. Three practical trajectories are likely:

• Workflow acceleration: Automated templates reduce repetitive setup work, letting teams focus on decision-making.
• New product surfaces: Companies may build integrations that export GenTabs to BI tools or task managers.
• Privacy-first controls: Session-scoped permissions will become table stakes for enterprise adoption.

Examples of near-term developer opportunities:

• Build adapters that convert GenTab exports into Jira tasks or Notion pages.
• Design micro-UI components that can be embedded by GenTabs for niche workflows.
• Offer privacy-preserving transformation services that scrub PII before generation.

Timeline and expectations: Google will likely run multiple feedback rounds through 2026, followed by incremental feature additions such as team-sharing primitives, richer data connectors, and enterprise controls. Competitors will respond with their own session-driven assistants, raising the bar for integration and UX simplicity.

Actionable next steps for teams watching Disco:

• Run a pilot: identify a recurring multi-tab workflow and test Disco with a small group.
• Measure impact: track time saved and error reduction from exported GenTab outputs.
• Plan integrations: map where GenTab exports could land in existing pipelines.

In short, Disco demonstrates a pragmatic way forward: use AI to turn messy sessions into usable tools that accelerate work, while demanding new attention to permissions and exportability.

Google Labs’ Disco remains an experiment with limited access, but it already signals practical changes in how browsing and AI combine to create productivity tools. Expect iterative releases, expanded integrations, and enterprise-focused controls over the next 12–18 months as Google refines GenTabs and responds to tester feedback on usability and privacy, and watch how the disco experiment influences both browser design and AI-assisted workflows.

Current status: Disco is live for a restricted set of testers and continues to evolve based on usage data, privacy reviews, and developer feedback; next developments likely include broader access windows, export/connector improvements, and enterprise controls that make GenTabs viable for teams and products.