Sift

About

Sift is an AI-powered browser extension that analyzes marketplace product listings in seconds and tells you whether a product is worth buying — or worth skipping. One click runs a full background check on the listing: marketing claims, ingredient safety, active recalls, and seller trustworthiness. You get a composite trust score, a plain-English verdict, and a list of any red flags found.

How it works

When you're on a product listing page, click the Sift icon in your toolbar. The extension reads the listing URL and sends it to Sift's backend, which scrapes the full product page and runs it through a multi-stage AI analysis pipeline. Within seconds, you see a trust score, a verdict badge (Verified, Caution, Red Flag, or Unverified), and the top flags that affected the score. You can then open the full report in the Sift web app to read the complete breakdown.

What Sift checks

Marketing claim verification — Sift identifies every performance claim in the listing copy and cross-references it against published evidence. Phrases like "clinically proven," "dermatologist recommended," "boosts immunity," or "increases energy by 40%" are extracted and evaluated for substantiation. Claims that can't be verified, are exaggerated, or are actively misleading are flagged individually and contribute to a lower score.

Ingredient analysis — for supplements, cosmetics, food products, and personal care items, Sift extracts the full ingredient list from the listing or from its scrape of the product page. Each ingredient is evaluated for known safety concerns, regulatory status, and common reactions. Ingredients flagged by agencies like the FDA, EMA, or EFSA are highlighted in the report.

Recall and safety database checks — Sift queries public government safety databases in real time, including openFDA, the U.S. Consumer Product Safety Commission, and equivalent agencies, to check whether the product or its brand has any active or historical safety recalls, warnings, or enforcement actions on record. A red banner appears in the popup if an active recall is found.

Brand and seller assessment — Sift evaluates the brand behind the product, looking at how long it has been operating, any patterns of complaints or enforcement actions, whether the brand's own claims are consistent across listings, and how the seller's pricing compares to market norms. Brands with very short histories, unusual pricing, or complaint patterns receive lower trust scores.