RepsFinder

Upload a photo of any clothing item → AI identifies it → instantly find community QC photos from Weidian sellers.

How it works

📸 Upload photo
      ↓
🤖 Claude Vision identifies the item  (e.g. "Nike Air Force 1 Low White")
      ↓
✏️  Confirm (or edit) the item name
      ↓
🔍 Search Reddit + Google for community spreadsheets
      ↓
🧵 Extract Weidian itemIDs via regex
      ↓
🖼️  Display QC photo links → finderqc.com/product/Weidian/{itemID}

Stack

Layer	Tech
Framework	Next.js 14 (App Router)
Styling	Tailwind CSS 3
AI Vision	Anthropic claude-opus-4-6
Search	Reddit JSON API + Google Custom Search
QC photos	finderqc.com

Getting started

1. Clone & install

git clone https://github.com/xtasySyrup/repsFinder.git
cd repsFinder
npm install

2. Set up environment variables

cp .env.local.example .env.local

Edit .env.local:

ANTHROPIC_API_KEY=sk-ant-...          # Required — get at console.anthropic.com
GOOGLE_API_KEY=AIza...                # Optional — broadens results
GOOGLE_CX=...                         # Optional — your Google Custom Search engine ID

Reddit is queried anonymously via the public JSON API — no credentials needed.

3. Run

npm run dev

Open http://localhost:3000.

API routes

Route	Method	Description
/api/identify	POST	Sends base64 image to Claude vision, returns { itemName }
/api/search	POST	Queries Reddit + Google CSE for "{item} weidian spreadsheet", returns raw results
/api/qc	POST	Runs Weidian itemID regex over all results, returns { items[] } with finderqc URLs

Weidian itemID extraction

All three text sources (URL, title, snippet) from every search result are scanned with:

/itemID(?:=|%3D)(\d{6,12})/gi

This catches both plain and URL-encoded forms of itemID=1234567890.

Each found ID maps to:

• QC photos → https://finderqc.com/product/Weidian/{itemID}
• Seller listing → https://weidian.com/item.html?itemID={itemID}

Project structure

repsFinder/
├── app/
│   ├── api/
│   │   ├── identify/route.ts   ← Anthropic vision
│   │   ├── search/route.ts     ← Reddit + Google CSE
│   │   └── qc/route.ts         ← Weidian regex + finderqc formatter
│   ├── globals.css
│   ├── layout.tsx
│   └── page.tsx                ← Full client UI
├── .env.local.example
├── next.config.ts
├── tailwind.config.ts
└── tsconfig.json

Setting up Google Custom Search (optional)

1. Go to Programmable Search Engine
2. Create a new engine — set it to search the entire web
3. Copy the Search engine ID → GOOGLE_CX
4. Get an API key from Google Cloud Console → GOOGLE_API_KEY

Without Google CSE the app still works; Reddit alone surfaces most community spreadsheets.

License

MIT