Scan
Open the app. Capture a burst of photos. The AI separates the items, names them, dates them, values them. An evening with a spreadsheet becomes a single Scan.
A cabinet of curiosities
for what you own.
Photograph what's around you. The AI catalogs it in seconds — descriptions, values, categories. Ask your collection questions. Share lists with people you trust. And when someone describes what they need, you're matched. No listing step. No browse feed.
The room you're in
$723
in unused items sits in the average household. 65% of people say selling is too hard.
USPS · Mercari joint study, 2023
Five things, one cabinet.
Three of them happen alone, in your pocket. Two of them happen when you let someone in.
For you, first
Refresh
The system speaks in lavender when it has something to suggest. looks like one you already have You judge; the AI shows its work.
Ask
Type a question. Where's the spare HDMI? Do I already have a stand mixer? Your collection answers in lavender, pointing at the items it's referring to. 3 items referenced
For the people you let in
Share
Make a list out of your catalog — a moving-box inventory, a wishlist, what's in the attic. Hand it to a trusted circle. Households see everything by default; everyone else sees only what you've shared, on purpose.
Matched — not listed.
Someone describes what they're looking for — a replacement Polaroid battery, a lamp for a nightstand, a part for a 1998 push mower. The system reads the intent and matches it to items already in other people's cabinets. No browse feed. No listing form.
Why this works now
The economics caught up to the metaphor.
$0.00015/ item ~80items in onboarding ~$0.024to catalog a household
Per-item cataloging cost dropped to fractions of a cent in 2026 — Gemini Flash for Tier 1 cataloging, Pro for Tier 2 appraisal, on-device for everything that should stay on device. The unit economics that made this impossible in 2022 work now.
ADR-046 · two-tier pipeline · ADR-074 · cloud-default with on-device fallback
The ask
Two paths. Same product. Different speed.
Path A · behavioral validation
$500K–$1M seed
12 months. Small engineering team, organic growth. The goal is to answer one question: do people catalog, and do they accept offers on items they cataloged for personal use?
Path B · the Uber playbook
$5–10M uplift
Subsidized cataloging across 10+ market cohorts. Android. A real growth team. The bigger swing. Reserved for the investor who wants to fund the marketplace, not the validation.
Want to see the rest?
The long-form demos, the deck, and the financial summary live behind a code.
Request access