I testified before the Department of Justice in a federal fraud case. That experience is why SureVett exists.

My name is Nathan Hart, and trust verification has been the thread running through my entire career.

The Bitcoin wallet years

In 2012, I built one of the first Bitcoin browser extension wallets — BitWallet. This was the Wild West era of crypto: no regulations, no guardrails, and scammers around every corner. I watched firsthand as people got deceived because they had no way to verify what was real. The technology was exciting, but the gap between what was genuine and what was fraudulent was massive and growing.

The DOJ testimony

That work led me into the middle of a federal fraud investigation. I was called to testify before the Department of Justice in a case that crystallized something I'd been feeling for years: the systems we rely on to verify trust are fundamentally broken. Not just in crypto — everywhere people need to distinguish what's real from what isn't.

A decade at AWS

I spent the next ten years at Amazon Web Services, building infrastructure for companies training frontier AI models. That work is where I learned what “production-ready” actually means at enterprise scale — systems that millions of people depend on every day, with the operational discipline to match.

The Fakespot moment

On July 1, 2025, Mozilla shut down Fakespot. Mozilla reported 10 million+ users when it announced the closure in May 2025 — all of them lost their only real protection against fake reviews overnight. I looked at the alternatives: one was caught posting fake Reddit recommendations for itself. Another gave perfect scores to products with known fake reviews in my testing. The irony of fake review detectors you can't trust was too much.

Building SureVett

I combined the firsthand understanding of how fraud works from the DOJ experience with the infrastructure engineering discipline from AWS. The shipping product scores Amazon products with a statistical-signal engine that runs entirely in your browser. Six signals contribute to the grade: rating distribution (the J-curve test), verified purchase ratio, review velocity (reviews per day relative to product age), review count, seller trust (Amazon vs. FBA vs. third-party), and brand and product signals (brand-name patterns, Best Sellers Rank, answered questions, listing age).

Every grade shows its reasoning, signal by signal. SureVett surfaces signals, not certainty. AI-generated reviews in 2026 are good enough that anyone claiming perfect detection is selling you something. Deeper AI text analysis of individual reviews is on the roadmap — the side panel says so too — and I'd rather ship the explainable layer first than over-promise an LLM tier that isn't live yet.

SureVett launched on Chrome. An iPhone app, with a Safari extension for mobile browsing and a Share Sheet for checking products from the Amazon app, is next. Walmart, Target, and Best Buy support is also on the roadmap.

Why it's free

The core trust-grade analysis is free, and always will be. No ads, no tracking, no data sold to third parties, no affiliate links on Amazon. If SureVett ever makes money, it won't be by putting a paywall between you and what a product's reviews actually say.

If you got here from the blog, the comparison of every Fakespot alternative is here.