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A storefront explorer for Shopify stores.
翻译准备中
取自仓库 README
正在加载完整 README。 以下为已保存的摘录。
A storefront explorer. Paste a Shopify store URL and browse everything it
sells. Filter by brand, product type, tag, and price.
Live at trybackroom.vercel.app.
No account, no API key, no install.
1. Open the explorer.
2. Paste a store URL. allbirds.com and https://www.allbirds.com/ both work.
3. Filter the catalog in the sidebar.
Stores you load are saved in your browser, so they come back on your next visit.
That list stays on your device.
You'll need Node 20 or newer and
pnpm.
Then open http://localhost:3000.查看完整 README ↗A storefront explorer. Paste a Shopify store URL and browse everything it sells. Filter by brand, product type, tag, and price.
Live at trybackroom.vercel.app.
No account, no API key, no install.
allbirds.com and https://www.allbirds.com/ both work.Stores you load are saved in your browser, so they come back on your next visit. That list stays on your device.
You'll need Node 20 or newer and pnpm.
pnpm install
pnpm dev
Then open http://localhost:3000.
Most Shopify stores publish their entire product list at a public URL, usually
https://yourstore.com/products.json. No login, no API key. Backroom reads that
and shows you the result.
Three things make it more useful than opening the JSON yourself.
You get the whole catalog. Shopify serves that list a page at a time, 250 products each. Backroom keeps requesting pages until one comes back short, which is how it knows it has reached the end. A store with 5,000 products gives you all 5,000. There's a ceiling, and when a store reaches it the page says so instead of quietly showing you a partial list.
The messy metadata gets cleaned up. Stores fill these fields in wildly
different ways. Gymshark puts an entire category path in the product type field
(Womens>Apparel>Leggings>full_length) and its marketing tagline in the brand
field (Gymshark | Be a visionary.). Taken literally, that makes every product
its own filter option and repeats the tagline on every card. Backroom turns those
into "Full Length" and "Gymshark".
You don't wait on a blank screen. The page layout and filters appear immediately, then products stream in underneath. A large store takes a few seconds to fetch, and you can use the interface the whole time.
Only Shopify stores, for now. WooCommerce, BigCommerce, Magento, and custom storefronts aren't supported, and Backroom says so rather than showing you an empty list. Some large retailers block automated requests entirely. Backroom can't work around that and doesn't try.
No export, and no accounts. The FAQ on the home page covers the rest, including whether scraping a particular store is something you should do.
Deployed to Vercel. pnpm build produces a standard Next.js build with no
platform-specific output, so it runs on any Node host.
Changing the code? AGENTS.md has the project's conventions: commit style, linting, and which files are generated and shouldn't be hand-edited.