TabMind

Framing the Problem

Tech people drown in a "sea of tabs," losing flow, credentials, and deep-dive articles to digital noise.

We've all been there - 50 tabs open, zero memory of why, so we just hit 'Close All' and start from scratch. We’re building the bridge back to that lost work.

The Gap: Existing grouping tools demand cognitive load and manual consistency.

Idea Explanation

An intelligent visual roadmap that uses Gemini 3 Flash to transform a chaotic "Contextual Graveyard" into an organised, actionable knowledge graph.

We don't just group; we interrogate each tab - extracting summaries and flagging redundant, duplicate content.

The result: a clean, prioritised journey that filters out the "digital noise" and shows you exactly where your research connects, and most importantly, where to start. End the 'Tab Anxiety': you can finally close the chaos, knowing your research is structured, saved, and ready when you are.

Implementation

1. UX Journey

   1. User installs the TabMind extension from Google Chrome store.

   2. Uset signs in.

   3. User clicks 'Refresh' on the extension.

   4. User is taken to a full Mind Map view.

2. Retrieval Stage

   1. Data extraction: HTML is stripped and bundled into JSON Stash.

   2. Each tab entry includes URL, Content Summary, Title.

   3. RAG (Retrieval-Augmented Generation) pipeline:

      1. Chunking: Breaking long articles into "meaningful bites."

      2. Vector Embeddings: Turning text into math ([0.21, -0.33, ...]) to find semantic connections.

      3. Vector Database: Storing these vectors with metadata for high-precision retrieval. Stored in using Google Drive API.

3. Synthesis and Generation

   1. The backend bundles Top Relevant Chunks + User Persona + Structural Constrains to instruct Gemini 3 Flash to map he ties between fragmented tabs.

   2. Gemini 3 Flash performs a semantic audit: clustering overlapping themes, filters out biased noise.

   3. A Specificity Slider on the frontend toggles the depth of the JSON tree, allowing users to collapse broad categories or expand granular sub-clusters on demand.

   4. The final hierarchical map is pushed to the Google Drive API, transforming a fleeting "Contextual Graveyard" into a permanent, shareable research asset.

Front end

• Mind Map View

  A full-screen visualisation of the entire tab ecosystem, using collapsible arrows to expand clusters and navigate the research journey from high-level pillars to granular details.

• Contextual Side-Panel

  An active sidebar that syncs with your current tab, displaying its exact position in the reading path (e.g., Step 3 of 10), estimated reading time.

• Smart Export One-click Export to PDF
  generates a clean, structured version of the full research mindmap and categorised reading list for offline review.

Challenges

We were struggling with the coding part, GitHub, but with the help of Gemini, we were able to finally develop the extensions and the TabMind.

Accomplishments

1. Managing time for the product ideation part.

2. Quick distribution of tasks in the group.

3. Balancing workload evenly across technical/ideas/design parts.

Next Steps

In the Google Ecosystem

1. Cross-device continuity: laptop to phone

2. Dynamic switching between Specialised LLMs (Academic, Dev, etc)

3. Collaborative mindmap

4. Multimodal expansion: image analysis & automated visual mood boards

Integration

1. Mobile integration across iOS and Android mobile web browsers

2. Extension to all major desktop engines

Additional Resources

Slides: https://www.canva.com/design/DAHDR-rGoQs/v29NqYCxp3tH6rAWjUoPsw/view?utm_content=DAHDR-rGoQs&utm_campaign=designshare&utm_medium=link2&utm_source=uniquelinks&utlId=hafe2860887