Paste any Shopify store. See what your customers are seeing.
No login, no install, nothing to authorise. Reads only what your storefront already publishes.
You have someone for ads, someone for the theme, someone for email. Nothing owns the product data all three of them read from. It accumulates other people's markup, loses its shipping weights, and drifts. Nothing checks it before customers see it.
Every bar is live right now. Nothing here was cleaned up. The oldest defect we found had been sitting on a published product page since 2018 — 8.3 years.
Volume doesn't explain the shape. Between 2024 and 2026 the number of products created grew 1.6×. The number carrying markup grew 4.9×.
The bottom rows are the proof that products get rewritten in place.
I run a small direct-trade tea and spice brand. One afternoon I found a tea in my own store with a shipping weight of 225 pounds. A hundred kilos. A seventy-dollar box of loose leaf. It had been live for months. Nobody had noticed, including me.
So I wrote something to read my catalogue properly, then ran it across 633 stores to find out whether I was unusual. I wasn't. Fewer than a fifth came back clean.
Along the way it turned out you can date the damage. Markup from Word, Slack, Wix, Google Sheets and now the AI chat tools gets pasted into product descriptions and stays there. It hides behind a page that still looks normal, so nobody catches it. The oldest one I found had been live since 2018.
— Arun, Living Roots USA
| Stores carrying at least one live defect | 81% |
| Expensive products with almost no description | 43% |
| Product photos too small to render sharply | 36% |
| Serving ChatGPT or Claude markup to customers | 1 in 3 |
| Hero images in more than one shape | 25% |
| Colour variants with no photo of their own | 22% |
| Tags spelled more than one way | 19% |
| Word markup still embedded in descriptions | 1 in 4 |
| Catalogs that lost their shipping weights halfway | 12% |
| Longest a defect had been sitting live, undetected | 8.3 yrs |
Two halves. The scan is free and needs no install. It reads only what your storefront already publishes.
For the repair, you pick how it lands: a bulk-edit CSV you import yourself, a list of exact edits for your admin, or temporary write access so I do it. The first two need no permissions at all. Either way you see every diff before anything changes.
It strips the foreign markup without altering a single word of your copy. Verified word for word, not by eye.
Here is one of mine, before and after. Same words, both times.
<div data-test-render-count="1"><div class="group"> <div class="contents"><div data-is-streaming="false" class="group relative pb-3"><div class="font-claude-response relative leading-[1.65rem]"><div><div class="standard-markdown grid-cols-1 grid gap-3"> <p class="font-claude-response-body break-words whitespace-normal leading-[1.7]">Some teas stand on their own. These three are better together</p>
<p>Some teas stand on their own. These three are better together</p>
That was 22 products in about twenty minutes. Descriptions came out 40–75% smaller and read identically on the page.
That's one of three things it checks:
It does not claim to lift your conversion rate. I looked for that and couldn't prove it, so I'm not going to say it.
The same product data is reused across your storefront, your collections, your shopping feeds, your campaigns and your fulfilment systems. A defect introduced once can travel through all of them.
And a product with broken data stays live. It still ranks. It still gets advertised. It still gets bought.
It also compounds. Among stores with fewer than 50 products, 55% had something wrong. Among stores with more than 500, it was 98%.
Catalog mistakes are common. The question is who finds them first: your team, or your customer.
The scan needs no install, so an entire client list can be read at once. Send me the client list you already publish. I'll scan every storefront on it and send back a QA report across the portfolio, then fix the first ten defects on any one of them, free, so you can see what a fix looks like before anyone touches a live store.