Wine cellar tracker with drink-by alerts

ConsumerMobile1 monthReact NativeExpoSupabaseRevenueCat

The idea

CellarTracker proved 800k people will voluntarily pay to track wine, on an interface that has not meaningfully changed since 2004: desktop grids, manual data entry, no urgency model. Meanwhile the actual failure mode of collecting is simple: bottles quietly slide past their peak because nothing tells you a window is closing.

The wedge is two mobile-native moves the incumbent cannot make easily. Entry by label photo instead of forms, and a Drink Soon queue driven by drink-window data, culminating in one monthly notification: open these next. That reframes the product from inventory (a chore) to not wasting the wine you already bought (a fear).

Who pays

The buyer holds 50 to 500 bottles, has watched at least one prized bottle turn out tired, and already demonstrates willingness to pay via CellarTracker’s voluntary model. Price at 30 dollars a year via RevenueCat with a free 25-bottle tier. Distribution: wine subreddits and forums where CellarTracker UI complaints are a genre of their own, wine YouTube, and a CSV importer that makes switching a ten-minute job.

MVP scope

  • Label photo entry with OCR lookup and top-3 candidate confirmation
  • Drink windows auto-suggested by style and vintage, user-overridable
  • Drink Soon queue plus monthly open-these-next notification
  • Rack locations, consumption log with quick ratings, cellar stats
  • CellarTracker CSV import
  • Skip for v1: market valuations and price tracking, social features, sommelier recommendations, barcode scanning
// Build prompt — paste into Cursor / Claude Code / Lovable / Bolt
Build Peak, a mobile-first React Native (Expo) cellar tracker whose organising principle is the drink window, not the spreadsheet grid.

Adding a bottle: photograph the label; run OCR on-device, then match producer, cuvee, and vintage against a wine reference lookup, presenting the top 3 candidates to confirm. On match, prefill region, grape, and a default drink window derived from vintage plus typical ageing curves per style (store a style-to-window table: e.g. village Burgundy 3-10 years, Napa Cab 5-20). User can override the window. Manual entry fallback with the same fields. Record quantity, purchase price, and rack location (user-defined zones like Rack A row 3, or fridge shelf).

The core screen is Drink Soon: bottles ranked by window urgency in three bands: drink now, entering peak, hold. A monthly push notification lists 3-6 bottles to open next, biased toward bottles the user has multiples of.

Consumption flow: mark a bottle drunk, add an optional 10-second note and 1-5 rating; quantity decrements and the note attaches to future bottles of the same wine.

Screens: Cellar (search and filters by region/style/location), Add Bottle, Bottle detail with window timeline graphic, Drink Soon, Stats (bottles by region, cellar value).

Stack: Supabase for data and label image storage, RevenueCat: free to 25 bottles, 30 dollars per year unlimited. Include CellarTracker CSV import to poach incumbent users.

// More Consumer ideas