Live Demo: feedbacklens.vercel.app
Product teams and startups collect vast amounts of customer feedback from surveys, reviews, and support chats — but most of it sits unused in spreadsheets.
Manually reading, tagging, and analyzing thousands of comments is:
- ❌ Slow
- ❌ Prone to human bias
- ❌ Difficult to scale
Feedback Lens was built to automate this workflow using GenAI.
-
✨ AI-Powered Analysis
Uses Google’s Gemini model (via Genkit) to perform fine-grained sentiment analysis and extract topics for each comment. -
📊 Interactive Dashboard
Visualize sentiment trends, explore topics like UI/UX, Pricing, or Performance, and filter insights dynamically. -
📁 Feedback History
Save each analysis to Firestore, with a "History" page to access past reports. -
🧮 Comparison Mode
Select two reports to generate a Comparison Dashboard with:- 🟢/🔴 KPI Deltas
- 📈 Overlaid trend charts
- 📌 Highlighted topic shifts
| Layer | Tech |
|---|---|
| Frontend | Next.js 14, React, TypeScript, Tailwind CSS |
| Backend / AI | Google Genkit (Gemini model) |
| Database | Google Firestore |
| Deployment | Vercel |
Challenge: Hitting API rate limits during large file analysis
- One AI call per row
- Scaled poorly
- Frequently timed out
- Batch processing: Grouped 15–20 rows per AI call
- Reduced API calls by 90%
- Improved speed and reliability
- 🧪 Product Managers
- 📈 Business Analysts
- 📣 User Research Teams
- 🚀 Hackathon & Startup Launches
Developed by @bandya2003
A full-stack AI project, crafted for clarity and strategy.
“From CSV to Insight – in under 60 seconds.”