PolyGrub
February 22, 2026
·
1 min read
PolyGrub lets you upload a photo of an ingredient list and uses AI to flag noteworthy ingredients - dairy, meat, common allergens, or anything else you configure - with confidence scores and category grouping.
Key Features
- Photo scanning - upload or capture an ingredient label photo, resized client-side before upload
- AI ingredient analysis - a Supabase Edge Function analyzes photos and returns flagged ingredients with confidence percentages
- Custom categories - create ingredient categories (e.g. “Dairy”, “Allergens”) with color coding
- Custom ingredients - define specific ingredients to watch for within each category
- Real-time updates - analysis status and results stream in live via Supabase Realtime
- Re-analyze - trigger a new analysis run on any previously uploaded photo
Tech Stack
| Layer | Technology |
|---|---|
| Framework | Next.js 16 (App Router) |
| Language | TypeScript 5 (strict mode) |
| UI | React 19, Tailwind CSS 4, Heroicons 2 |
| Auth & Database | Supabase (Auth, Postgres, Storage, Realtime) |
| AI Analysis | Supabase Edge Function (triggered by DB insert) |
How It Works
Photos are resized client-side, uploaded to Supabase Storage, and a database trigger kicks off an Edge Function for AI analysis. Results stream back to the client in real-time via Supabase Postgres Changes, so you see ingredient matches appear as they’re identified.