Skip to content
azizu06Public

About

AI assistant that turns a photo of a used item into a priced, ready-to-post marketplace listing.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Latest commit

 

History

1,006 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SnapList

SnapList is an iPhone app that turns a few photos of something you own into a priced, ready-to-post resale listing. AI identifies the item, checks real eBay sales for a price, and writes the listing for you to edit.

Built for RevenueCat Shipaton 2026, Next Gen track

Built for RevenueCat Shipaton 2026 · Next Gen track · Native iPhone app (SwiftUI) with a TypeScript backend

Native iPhone capture: real Jordan 3 shoes in SnapList's camera.   Native To list: a backpack and Logitech mouse are ready to review, both with real thumbnails.   Native Jordan 3 review: $155 price, three sold matches from $100 to $180, and editable listing details.

Capture → To list → review. Stills from the native iPhone demo recording. Screenshot notes and source timestamps.

SnapList demo

Watch the demo video

Run it

The backend needs Node.js 22.19.0 and pnpm 10. From a fresh clone:

nvm use                           # picks Node 22.19.0 from .nvmrc
pnpm install --frozen-lockfile
cp .env.example .env.local        # fill in the keys you need
pnpm test                         # unit tests (Vitest)
pnpm dev                          # start the Next.js API server

For the iPhone app, open ios/SnapList.xcodeproj in Xcode 26.5, pick the SnapList scheme and an iPhone simulator. The developer guide covers local Supabase and the full set of checks.

The original overview, how it works, and the Pro subscription details continue below.

What it does

Turn the stuff you own into listings worth posting.

Snap a few photos, add a quick voice note, and get an editable resale listing with a starting price.
Spend less time researching and writing. Keep the final say on what goes out.

Scout, SnapList’s camera mascot, scanning a barcode on a cardboard parcel in a looping animation.

From photo to flip

  1. Snap it. Take one to five photos. Add an optional voice note of up to forty-five seconds for details the camera might miss.
  2. Get a draft. AI identifies the item and writes a title, description, condition, and item details. Everything stays editable.
  3. Check the price. SnapList researches comparable eBay sales and shows the matches it can verify. If reliable matches are unavailable, you still get a clearly labeled starting estimate. Set your own price at any time.
  4. Keep going. Start the next item while earlier items process in parallel. To list keeps the ones waiting for review together, so you can work through several items without waiting between each one.
  5. Choose where it goes. Review and confirm a direct eBay publish. For Facebook Marketplace, Mercari, or Depop, take the prepared text and photos into the marketplace and finish posting there yourself.
  6. See your Flips. Finished items collect in a chronological view. An export marked prepared or shared stays distinct from an eBay publish; neither is a claim that the item sold.

Native eBay post success: Jordan 3 shoes are marked Live on eBay.   Native Flips: the mouse and posted Jordan 3 shoes appear with real photos, October 1 dates, and the navigation dock.   Native Facebook Marketplace sharing drawer: a Logitech mouse at $44.95, two photos, and text/photo handoff steps; all destinations show Not started.

Posted to eBay → Flips. Facebook Marketplace uses a guided handoff; this screen shows Not started.

The first usable listing comes before signup or a paywall. Try the result on your own item before deciding to keep listing.

RevenueCat powers SnapList Pro

Scout meets you at a packing counter when you want more AI listings. The plan appears on a taped packing slip: price, renewal terms, Subscribe, and Restore, with Terms and Privacy close at hand.

Native sandbox subscription confirmation: Scout celebrates SnapList Pro is on beside monthly plan terms.
SnapList Pro is on — Apple sandbox demonstration. Screenshot notes.

Free first item SnapList Pro
One complete AI listing, including one guided identity correction for the same photos. Review and edit the result before committing to a subscription. Continue with new AI items through a monthly allowance. The second complete AI item run requires Pro.

RevenueCat handles the Apple subscription flow: the iOS app loads the configured offering and localized product price, purchases through its SDK, and supports restoring purchases. The same integration connects the subscription to the signed-in SnapList account.

The server confirms access. RevenueCat subscription events update the verified Pro entitlement and billing period. A successful purchase screen alone cannot unlock AI credits. The app has confirming and pending states, and can resume the saved item once access is verified.

Credits count useful results. A credit settles when an editable, priced draft is saved. Technical retries and recovery reuse it; a failure before that point restores it. Pro’s monthly allowance is configurable while real usage costs are measured.

The integration is implemented and covered by purchase/restore fixtures and server contract tests.

Built for the parts that matter

  • An iPhone experience, not a web form. SwiftUI capture, optional voice context, compact editable fields, and Scout keep the work close to the item in your hand.
  • Work that survives interruption. The TypeScript backend saves progress between identification, pricing, and writing. A recovered job continues from saved work instead of starting every AI call again.
  • Evidence you can question. Sold results pass an item matcher before they support a price. Confidence comes from identification and pricing evidence. When research falls short, the app says so.
  • Your edits reach the marketplace. One shared price rule carries your chosen price into eBay and export packs. Publish confirmation is protected against duplicate requests.
  • Private by design. Supabase database policies isolate each seller's items; photos and voice notes use private storage. Guest access is tied to Apple's App Attest, with an encrypted result recoverable for 24 hours.

Explore the implementation, tests, and measured benchmarks.

See it in action

Watch the demo video or read the Devpost write-up. This repository is the Shipaton showcase.

To explore locally, open ios/SnapList.xcodeproj in Xcode and follow the iOS setup guide. The developer guide covers the API, local Supabase, and verification commands.

For builders

Developer setup & architecture · Product requirements · Documentation index · Screenshot provenance

Apache License 2.0

About

AI assistant that turns a photo of a used item into a priced, ready-to-post marketplace listing.

Topics

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages