- Before
- Pricing, descriptions and listing were done by hand, one item at a time
- What we built
- Drop in photos and several AIs research the market, write the price, title and description, then list the item and confirm the sale. Rolled out to a second channel (BASE) as well
- What changed
- Giving the listing order takes about a minute, whether it is 30 items or 70. People only handle the photography: 56–78 items a day (7 working hours)
Screens
Screens are from the trial edition distributed to other companies, with test photos. Left: right after dropping in photos. Right: price and description done. The interface is in Japanese.
We list with this app in our own store, too
The founder’s own eBay store (trading as ALMOST WIPEOUT — vintage clothing and retro games) lists with the same app. The listing automation we ran for that store — AI reads the photos, researches the market, writes the description and lists the item — is where this app came from. Our first client engagement also started from a conversation about it.
- Period
- Feb – Oct 6 (2026)
- Items listed with this app
- about 1,300
- Sales from those items
- $7,000–8,000
Figures from the founder’s own store, as of 2026-10-06. Prices and the sold count on the screenshots are blurred.
StackElectron / Claude, GPT, Gemini, Grok (market research and judgment) / Yahoo! Shopping Store API / BASE / Claude Cowork