PixelVision · iOS & Android
SnapPrice AI
Snap it. Price it.
Point your camera at any used item and get a three-tier resale price - Quick Sell, List For, Optimistic - blended from 17 live marketplaces in seconds.
iOS + Android · Gemini Vision · 17 marketplaces
One number is a guess.
Every other resale tool hands you a single "market price" and calls it done - but the right number depends entirely on how fast you need to sell. SnapPrice reads the item, checks the live market, and gives you the whole range: a floor for a fast sale, a data-backed list price, and a ceiling for the patient. The call stays yours.
The pipeline
From one photo to three prices
Gemini Vision names the item; a live marketplace blend prices it; the result is a range, not a guess.
The capture
Snap, confirm, priced
Point your camera at any item - or pick a photo from your library. Gemini Vision reads category, brand, model and condition straight from the pixels, no typing required.
A pre-filled card shows what the AI detected. Add or fix a detail for a tighter number - for most items you just tap through.
SnapPrice queries 17 marketplaces in parallel, blends them with the AI valuation, and returns Quick Sell / List For / Optimistic - plus a plain-English appraiser's note.
The output
Three prices, not one
Most resale apps give you a single "market price". SnapPrice gives you a range - because the right price depends entirely on how fast you need to sell.
Price at or below this for a fast sale, often within 24h. Useful when you need cash quickly or the item is hard to store.
The data-backed sweet spot. Sets realistic expectations and attracts serious buyers without leaving money on the table.
The ceiling for patient sellers in markets where demand is high. Best for rare items, limited editions, or vintage pieces.
Market data
17 marketplaces, one lookup
The app
Built for how people actually sell
Under the hood
Three layers, one tap
The pipeline is deeply complex - a vision model, a parallel market lookup, a cache layer, currency conversion, condition normalization. The user sees a camera button and a price.
Identifies category, brand, model and condition from the photo - electronics, fashion, furniture, instruments, sporting goods and more.
Parallel queries across 17 providers, filtered by country, condition and listing age, then normalized and blended with the AI valuation by a weighted average.
A 7-day cache keyed to brand + model + country + condition. Repeat lookups return instantly; the hit rate on popular items tops 80%, keeping API costs viable at scale.
What we learned
The right abstraction wins
The hardest decision was how much to show the user. Everything underneath - the vision model, the market lookup, the cache, the currency math - exists to serve one screen: a photo in, a price out. Every engineering call was made in service of that simplicity.