BrainLens is a multimodal research copilot powered by Gemini 3, built for the Gemini 3 Hackathon.
It turns dense academic files into a visual and navigable interface — not just a learning tool!
BrainLens helps you verify claims, scan evidence, and turn ideas into experiments fast.
-
Intelligent Circuiting: Get high value informations.
-
A-HA Mode: Get powerfull insights and deep-dive into then to generate formal hypotheses and experimental protocols.
-
Evidence Visualizer: Automatic extraction and relevance analysis of figures and tables.
-
Consistency Checker: Verifies claims against the source material with confidence scores and page citations.
Academic PDFs are hard to:
- Navigate quickly (key points are scattered across pages)
- Trust at a glance (claims often sound strong, but are they supported?)
- Turn into action (insights stay “cool ideas” instead of testable plans)
BrainLens uses Gemini 3 to power a multimodal, structured reasoning pipeline over academic PDFs:
-
Briefing (source-linked map)
A high-level paper map where each item links back to the page context — fast visualization without losing traceability. -
Consistency (claim checking)
Extracts important claims and checks them against paper context, labeling each one with High / Medium / Low confidence. -
Evidence Cards (figures & tables)
Surfaces key figures/tables, makes them readable, and explains why they matter — a clean “scan the evidence” view. -
Insights (reasoning-first)
Generates structured insight cards:- Alternative Lenses
- Hidden Assumptions
- Cross-Domain Links
- Actionable Next Steps
-
A-HA Mode (insight → experiment plan)
From any Insight card, click “A-ha!” to instantly generate a research experiment plan:- Research question
- Testable hypotheses
- Variables & controls
- Metrics & expected effects
- Threats to validity
Other experiences:
- Live Explanation (talk to the paper; adaptive depth/pacing)
- Podcast Mode (listen to the paper like an episode)
- Not just summarization: outputs are structured for analysis and decision-making
- Traceability: ideas are tied back to the source context/pages
- Reliability signals: claim-to-context checking with confidence labeling
- Actionability: insights become testable experiment plans in one click
- Upload a dense academic PDF
- Open Briefing to understand structure + jump to sources
- Check Consistency to validate key claims
- Scan Evidence Cards to see figures/tables clearly
- Explore Insights, then click A-ha! to generate an experiment plan
- (Optional) Use Live Explanation or Podcast Mode
- React + TypeScript
- Vite
- Gemini 3 API (multimodal generation; text + audio/TTS depending on enabled features)
- UI: Tailwind classes + Lucide icons
- A Gemini API Key from Google AI Studio
-
Clone and Install:
npm install
-
Environment Setup: Create a
.env.localfile in the root and add your key:GEMINI_API_KEY=your_gemini_api_key_here
-
Run Development Server:
npm run dev
The project has been refactored into a modular, scalable structure:
src/
├── components/
├── views/
├── services/
├── pipeline/
├── prompts/
├── constants.ts
└── types.ts
- Grounding vs. speed: balancing “fast demo UX” with traceability and structured outputs.
- Consistency checking reliability: designing prompts/logic that flag weak support without becoming overly conservative.
- Evidence readability: making figures/tables scannable in a way that feels visual and truly useful (not just extracted).
- Hackathon constraints: prioritizing an end-to-end experience quickly, while keeping the pipeline modular.
- Backend proxy for Gemini calls (key protection + caching + rate limiting + retries)
- Stronger grounding: more explicit citation snippets + tighter page linking for every claim/evidence item
- Richer evidence parsing: improved figure/table interpretation and cross-references inside the paper
- Insight quality upgrades: better ranking, deduplication, and “related insights” clustering
- Export & collaboration: export a structured report (PDF/MD) and shareable “experiment plans”
- Multi-PDF comparison: compare multiple PDFs to surface convergences, contradictions, and complementary evidence across papers
Hackathon prototype / demo project.
Built with Gemini 3 and a science-first philosophy: verify evidence, generate insights, and turn them into action.