Shop

Reviews, checked the way evidence should be.

Fake reviews are a documented problem: paid five-star blocks, bot-written testimonials, and listings where every “verified purchase” traces back to the same handful of accounts. Shop runs the same evidence-and-independence engine behind ASURIQ on reviews and listings themselves — not just on what an AI says about them.

What it checks

Not sentiment. Evidence.

Shop doesn’t score how positive a review sounds. It checks whether the evidence behind a listing’s reputation holds up — and discloses plainly when it doesn’t.

Review fingerprinting
SCRVNR-style writing-pattern analysis flags review clusters that read like one author, not many — the same fingerprinting technique built to score generated text in milliseconds, pointed at listings instead of chat.
Ownership, not just star count
A product with 4,000 five-star reviews from 4,000 independent accounts is a different situation from the same count funneled through one seller network. Independence ships as fact, the same way it does everywhere else on ASURIQ.
Marketplace certification
A listing that checks out can carry a certificate — the same resolving /v/[id] record every other ASURIQ verdict produces. A fake cannot resolve one.
Your data stays yours
Prompts stay with your provider. We see analysis metadata only.
Keys never stored
One-way hash for authentication. Your credentials pass through. Never persist.
~2 second responses
Single-model cognitive tools return in about 2 seconds.
37 of 100 flagged
We ran 100 ChatGPT answers through verification. See the study →

Currently in friends-and-family beta. Built on peer-reviewed cognitive architecture.

Baars — Global Workspace TheoryACT-R — Memory Decay ModelWang et al. 2025 — Silent AgreementLi et al. EMNLP 2024 — Sparse Debate

Know before you buy.

Friends-and-family beta · get access, try it on a real listing