尚未检测到更新。GitHub 发布版本后会显示在这里。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证正在加载完整 README。 以下为已保存的摘录。
Internal Sticker Peak tool for turning shelf inventory into eBay listings. Two stages:
1. Scan (/scan) — mobile-first. Photos, UPC barcode scan, quantity, multi-pack info, expiration, shelf location. Saves straight to the database as soon as each item is submitted, no local-only state to lose.
2. Review (/review) — any device, desktop included. Edit title/description/price/category/condition, mark items ready, then download a File Exchange CSV for Seller Hub.查看完整 README ↗Internal Sticker Peak tool for turning shelf inventory into eBay listings. Two stages:
/scan) — mobile-first. Photos, UPC barcode scan, quantity, multi-pack info, expiration, shelf location. Saves straight to the database as soon as each item is submitted, no local-only state to lose./review) — any device, desktop included. Edit title/description/price/category/condition, mark items ready, then download a File Exchange CSV for Seller Hub.It's a normal website, not a native app: the same URL works in Safari on iPhone and Chrome on Android for scanning, and in any desktop browser for review/export.
Deployed and running on Vercel with a live Supabase database. Scan flow, review/export flow, UPC lookup, Claude-drafted titles/descriptions, and a self-learning eBay category ID lookup (see below) are all wired up. Not yet done:
No eBay API access means no Taxonomy API, and this app doesn't scrape eBay's site at runtime (against their terms, fragile). Instead, category IDs are looked up once per product type and saved as a CategoryRule (keyword → category ID) in the database. The review page matches an item's title against saved keywords and auto-suggests the category; anything set manually can be saved as a new rule with one click ("Remember this category"). See references/ebay-category-ids.md for the running log and how to find a new ID when one's needed.
npm install
Copy .env.example to .env.local and fill in:
DATABASE_URL / DIRECT_URL — from a Supabase project (Project Settings → Database → Connection string). Use the pooled connection (port 6543) for DATABASE_URL, the direct connection (port 5432) for DIRECT_URL.CLOUDINARY_* — from your Cloudinary dashboard (Settings → Access Keys).ANTHROPIC_API_KEY — from console.anthropic.com (a separate login from claude.ai). Powers the "Generate AI draft" button on the review page.UPC_LOOKUP_API_KEY — not needed yet, UPC lookups use UPCitemdb's keyless trial tier (100/day) until volume requires a paid key.Push the schema to your Supabase database:
npx prisma db push
(We use db push instead of migrate dev here — Supabase's pooled connection doesn't allow Prisma to create the shadow database migrate dev needs. prisma/migrations/ still has the initial migration for reference, but schema changes since then are applied directly.)
Run the app:
npm run dev
Deploy to Vercel, pointed at this folder. Add the same env vars from .env.local as Vercel project environment variables. Turn on Vercel Deployment Protection (Project Settings → Deployment Protection → password) so the site isn't a public URL with no lock on it, since it holds inventory and pricing data. Note: @zxing/library currently warns it wants Node 24, Vercel's Node runtime version can be set in Project Settings if the build complains.
src/lib/csv.ts generates a minimal "Add" format (Action, SKU, Category, Title, Description, PicURL, Quantity, StartPrice, ConditionID). Before your first real upload, download an actual template CSV from Seller Hub for your category and compare headers, File Exchange's expected columns can vary slightly by category.