Type what you ate — get instant calories and macros via Perplexity Sonar AI. No barcode scanning, no database hunting.

Calorie apps make you search databases, scan barcodes, and weigh portions. You just want to log "rice and beans with avocado" and move on.

Traditional trackers optimize for precision theater, not daily adherence.

PWA calorie tracker: natural-language food logging, frequent meals, push notifications, offline queue. Live at eat-note.vercel.app.

Input: "grilled chicken salad with olive oil dressing"
→ calories, protein, carbs, fat estimated instantly

git clone https://github.com/irachrist1/eatnotes.git && cd eatnotes
npm install
npx convex dev    # terminal 1
npm run dev       # terminal 2

Deploy: vercel --prod

• Natural language in, macros out. /api/analyze-food calls Perplexity Sonar — no manual nutrient lookup.
• Clerk auth + Convex. User data persisted with typed schema; offline queue syncs when back online.
• PWA service workers. Installable on phone; works offline for logging, syncs later.
• Frequent meals shortcut. One-tap re-log of regular items — reduces friction for repeat breakfasts.
• Web push cron. Reminder notifications via scheduled jobs — nudge without nagging.
• Perplexity, not a food DB. Estimates from language understanding — fast enough to actually use daily.

Built by Christian Tonny