Vera

Vera is an iOS app that turns a bottle of water into a full purity report. The user opens the app, scans the barcode or snaps a photo of the label, and within seconds gets a 0 to 100 Purity Index along with a breakdown of what is in the water.

Each scan surfaces pH, TDS, mineral content (calcium, magnesium, potassium, sodium), and contaminant levels for fluoride, nitrates, lead, chlorine, PFAS, and microplastics. Anything flagged gets a plain-English explanation so a non-chemist can read it and decide whether to drink it.

The product is aimed at health-conscious consumers who already care about what they put in their bodies but have been flying blind on the water side. The pitch is simple: one scan, full transparency.

The hard part was not the app. It was the data. There is no clean global database of bottled water composition, so we built one from scratch. We pulled from public lab reports, municipal water quality records, regulatory filings, and the disclosures brands publish on their own sites. Every source is structured differently and updated on a different cadence.

We built an ingestion pipeline that normalises all of that into a single schema per product, with provenance attached to every field. When a brand updates their report, the pipeline re-ingests it and the app reflects the change without a release.

The scanning layer combines barcode lookup with vision-based label recognition for products that are not in any barcode registry, which is a surprising number of regional brands. When a barcode lookup fails, the LLM reads the label image, extracts the brand and product line, and matches it against our database.

The Purity Index itself is a weighted score across mineral quality, contaminant levels, source integrity, and testing transparency. We tuned the weights over a few rounds of user testing so the scores matched the intuition of someone who actually cares about this stuff.

We shipped Vera to the App Store, and the app now covers thousands of bottled water products with the data pipeline continuing to expand coverage automatically. Users can walk into any supermarket, scan a shelf, and get an instant read on what they are about to buy.

The bigger result is that a category that has been deliberately opaque now has a consumer-facing layer of accountability. Brands that publish full lab data score better. Brands that do not, do not.