★ A Public-Service Bulletin for Sellers ★
Has your product passed inspection?
"See your product exactly as the robot shoppers see it — one free audit, backed by the unerring accuracy of BAYESIAN SCIENCE!"
★ ★ ★
AUTO-DETECTS AMAZON · WALMART · TARGET · BEST BUY. FAIR-USE LIMITS APPLY — THE BUREAU IS BUSY.
0 = unsupported claims · 100 = fully evidence-backed
What the robot shopper sees–
Evidence observed ✓
Evidence missing ✗
How to pass inspection
Robot agents read this score live
This inspection runs on the same evidence engine agents query in real time through GoBuy's MCP server — tools: check_product_trust, compare_products.
Score-Movement Telegram (coming soon)
Leave your address — the Bureau wires you when your score moves and how you stack up against competitors.
★ How the Bureau works ★
Present the product
Drop in an Amazon, Walmart, Target or Best Buy listing URL — or a bare product ID with the retailer selected.
The evidence engine runs
We query the same marketplace-evidence engine AI agents consume live through GoBuy's MCP server.
Receive your report
Evidence Score, the agent's own verdict, observed vs missing signals, and concrete steps to pass inspection.
★ Questions from the trade
What is an Evidence Score?
A 0–100 measure of how well a product's marketplace evidence supports the listing's claims. It's computed from independently observable signals — seller history, verified-purchase reviews, Q&A sentiment, review integrity, price history — and normalized so AI agents can act on it. Higher means more evidence an agent can trust.
Which retailers are supported?
Amazon, Walmart, Target and Best Buy at launch. Paste a product URL and the retailer is detected automatically; for a bare product ID, pick the retailer from the dropdown.
How is this different from a seller rating?
Seller ratings are one marketplace's self-reported average. The Evidence Score aggregates multiple verifiable signals about the product itself, weighs only what an independent observer could confirm, and shows exactly which evidence is missing. It's built for machine consumption — AI agents read it live via MCP.
Is the inspection really free?
Yes — free within fair-use rate limits. Optional score alerts and competitor benchmarks are coming soon; leave your email above to get notified.
How often does the data refresh?
The evidence engine refreshes continuously upstream. Audits are cached for 15 minutes, and every result shows its freshness state (fresh or stale) plus when it was computed, so you always know what you're looking at.
Can I use this in my own AI agents?
Yes — GoBuy exposes check_product_trust and compare_products tools over MCP at mcp.gobuy.ai/mcp. Point your agent at the server and it queries the same scores powering this inspection.