Step 1 · Browse
A listing shows up. Is there money in it?
Browse pulls live eBay listings through Flipz's own server — sneakers, power tools, camera lenses, ten categories on rotation, or whatever you type in. Endless scroll, real photos, real asking prices.
Every card is about to get the same treatment. Nothing is shown without a price behind it.
Listings come from eBay's Browse API through a Vercel function that holds the credentials server-side. No key ever ships to the browser.
Step 2 · Evidence
The comps gather — and every one says where it came from.
Three tiers, ranked by how much they can be trusted. Crew sales imported from real platform reports come first. Then eBay history Flipz snapshots itself, hourly, tagging listings that vanished as presumed sold. Last, a category ballpark that only speaks when nothing better exists.
Words are weighted by rarity, so “Panda” matters more than “Nike.” Sizes stay whole — 10.5 never splits into a 10 and a 5.
Step 3 · Estimate
A range, not a number. And a confidence that has to be earned.
Comps halve in weight every 90 days. Agreement between them raises confidence; thin evidence or a wide spread pulls it down. The factors multiply, so two comps that happen to agree can't fake a “Strong.”
Estimates built only from eBay history are capped at 5 out of 10. Category ballparks never read above 2. The interface won't let weak evidence look authoritative.
Step 4 · The math
Fees come off the top. What's left is what you'd actually clear.
The math runs at the real published rates: eBay's 13.6% plus $0.40 an order, the 8% sneaker rate over $150 that most calculators miss, Mercari's 10%, Poshmark's 20%. Likely sale, minus fees, minus what you'd pay — that's the profit on every card, before you drive out.
The same fee tables are unit-tested in both Flipz and the Flip Desk, so the number on the card is the number on your payout.
Step 5 · Verdict
Buy it or pass — in one glance. Then sort the whole feed by profit.
Here the seller wants $101 all-in against a likely $112 sale — after fees that's a loss, so it's a pass, and Flipz says by how much. Flip the feed to “Buy it” only, sort by profit, and the deals worth a message rise to the top.
Nothing here is an “AI opinion.” Every card can be opened to the comps behind it, with their match, age and weight.
Step 6 · Get paid, get ranked
You sell it. You import the report. You climb the board.
Drop the CSV from eBay, Mercari or Poshmark into the Data tab. Columns are detected for you. The sale counts toward your rank on the crew leaderboard — which ranks on imported revenue only — and becomes a comp everyone in your crew prices against. Typed numbers earn nothing, so the top of the board is earned, not claimed.
A platform export proves what something sold for. It never contains what you paid. Your cost basis stays in your private Flip Desk.
Seller 99.2% · 1,412 feedback