RecallLens AI

RecallLens is a recall-verification tool that checks the specific unit you're holding, not just the product line.

The problem

Most recall apps match on barcodes, which are identical across every unit of a product. But recalls usually apply to a specific lot, batch, or date range — hundreds or thousands of units off the same line can be fine while the rest are
recalled. A barcode match returns "no recall" and you've returned your unit; the correct answer is "your lot is recalled."

The agencies that issue recalls (FDA, CPSC, USDA, NHTSA) publish structured criteria — exact lot numbers, lot ranges, manufacturing date windows — but almost no consumer tool reads them.

What it does

1. Reads the package — Gemini extracts brand, product, variant, lot code, and dates from a single photo.
2. Matches deterministically — a rule-based engine compares what was read against each recall's criteria. AI reads the photo, but the safety call is auditable: "this product falls within recalled lots 4167–4169" with a full rule trace.
3. Returns one of four honest outcomes:
   • Potential match — identifiers fall within the recalled range
     
   • No match — brand/product match, but lot/date is outside the recall
   • More info needed — couldn't read enough; tells you exactly what to capture
   • Unable to verify — the notice lacks structured criteria
4. Links to the source — every result points to the original FDA/CPSC/USDA/NHTSA notice. The final call is yours.

Built With

• next.js
• typescript