Your product page. Somebody else's checkout.

We crawl the open web for shops selling your photographs, score what turns up against your catalogue, and put every one in front of a person before anything goes out in your name.

Book a scan of your store See a finding

Month to month Nothing filed without your approval

KK-0416 · found 4 minutes ago Needs your review

Yours from your catalogue

$95.00

northbaygoods.com/atlas-card-holder

Theirs captured just now

$27.40

atlas-wallets.shop/p/8841

Image match 94% Priced at 29% of yours Domain 11 days old

Example finding. Northbay Goods is a brand we made up, and the figures are illustrative.

A copycat doesn't just take one sale. It takes the next one too.

The lost order is the part you can count. These three cost more, and they compound quietly while you are busy running the business.

Monica R. Took six weeks to arrive and the metal is already bent. Nothing like the pictures on your website. ★☆☆☆☆

Jordan M. Ordered this and it never turned up. Total waste of money. ★☆☆☆☆

The reviews stop being yours

Someone buys a $27 copy and opens a bag of nothing like the photos. They have no idea they bought a fake. They leave the review under your name, and your support inbox gets the rest.

The sale still happens

Just not to you. Price is the tell. A copy undercuts far enough that a shopper doing one quick search finds a cheaper version of the exact photograph they were already looking at.

northbaygoods.com

northbaygoods-sale.shop

northbay-goods.store

nothbaygoods.com

northbaygoods.outlet

You pay for the click, they take the customer

Lookalike domains bid on your own brand name and rank directly underneath you. Every dollar you spend teaching people to search for you sends a slice of them somewhere else.

The crawler does the looking. A person does the deciding.

Matching photographs is the easy half. The half that matters is telling a counterfeit apart from a retailer you actually sell to, a review blog, and your own listing on a marketplace you forgot about.

Review queue 6 waiting

94 atlas-wallets.shop/p/8841Priced at 29% of yours, domain 11 days old Counterfeit

88 nb-goods-outlet.store/atlasYour photos, your copy, your logo Impersonator

61 carryandco.com/slim-holdersPriced at list, and on your seller list Authorised

44 everydaycarryreview.com/best-ofAn article with affiliate links, not a shop Not a shop

– marketplace.example/itm/44192Answered with a bot check instead of a listing Unreadable

The real review screen, running a made-up brand's data.

What the score is made of

Perceptual image matching against every photograph in your catalogue, then everything around the photograph: price against your list price, how old the domain is, whether the seller is on your authorised list, and whether the page is a shop at all.

And what it could not check

Price, because the page never showed one
Catalogue breadth, only one product page loaded
Domain age, the registry did not answer

Every finding carries this list. A reviewer who believes the price was compared, when it never was, is worse off than one who knows it wasn't.

Your call

Nothing leaves without you ticking it

Findings land in your dashboard with the evidence attached: both photographs, both prices, the live link, the page as we captured it and the minute we saw it. You approve or you decline. A decline teaches the matcher, so the same false positive does not come back next month.

Then our job

We write it, send it, and chase it

Approved findings become a takedown notice drafted against the right channel for that host, which is rarely the same twice. We file it, log what comes back, and keep the whole thread in one place so you are not reconstructing it from your inbox three months later.

One real scan, reported in full. Including the part that didn't work.

KopyKatch is in private beta, and we are not going to print a round number we have not earned. Here is the last complete run, against a public direct-to-consumer accessories brand.

product photographs read out of the catalogue

114

candidate pages found and analysed

opened as findings for a person to review

pages we could not read, and said so instead of calling them clean

Why that last number is on this page

Thirty-nine pages answered with a bot check instead of a listing. A tool that quietly counts those as nothing found hands you a clean report that is not true. Ours marks them unreadable and puts them in a list you can open, because "we could not get in" and "there is nothing there" are very different sentences.

What we are still fixing

Some of those 25 findings were wrong. Three were the brand's own listings on a checkout service it uses, two were review articles the matcher took for shops. We know because we read all 25 by hand. Tightening that is what the next release is.

What most brands do instead

All four of these work. They fail in different places, and the place they fail is usually the month you get busy.

	KopyKatch	A freelance VA	Someone in-house	A law firm
What it costs	Flat monthly, cancel any month	Hourly, and the hours grow	A salary, plus the hiring	By the hour, per notice
Who does the looking	A crawler on a schedule, checked by a person	A person searching by hand	A person searching by hand	Usually you, then you send it on
Time to the first finding	The week you connect your catalogue	Depends on their week	After hiring, then training	After the engagement letter
When your catalogue doubles	Nothing changes on your side	Roughly double the hours	You hire again	Roughly double the bill
Who files the takedown	We draft it, you approve it, we send it	Often nobody does	Your hire, learning as they go	They do, and they do it well

Where we can look, and where we have to knock

Most pages like this show a wall of platform logos and let you assume they all work the same way. They don't. Half the internet answers a crawler with a bot check, and pretending otherwise is how you end up with a clean report and a live counterfeit.

We read these directly

Open storefronts and the searches that lead to them. We fetch the page, capture it, pull the photographs and the price, and keep the evidence.

Shopify stores WooCommerce BigCommerce Squarespace Wix Standalone shops Web search Reverse image search

We go through the front door

The big marketplaces block crawlers, and we are not going to defeat that. We reach them the way they are meant to be reached: rights-owner and brand registry programmes, which is also the only channel that can actually get a listing removed.

Amazon eBay Etsy Alibaba AliExpress Temu Walmart TikTok Shop Meta ad library

Named as platforms, not as partners. Enrolment needs your trademark and takes a few weeks. We walk you through it.

What working with us is actually like

Small on purpose

You are not account 4,812. The person who built the matcher reads your findings, and answers you himself when one of them is wrong.

Month to month

No annual commitment, no lock-in. If a quarter goes quiet and you want to pause, pause.

It runs whether or not you remember

Scans are on a schedule. Copies go up on a Sunday just as happily as a Tuesday, and so does the crawler.

Evidence, kept

Every finding stores the page as we saw it, both photographs, the price and a timestamp. A notice with a capture behind it goes better than one without.

A weekly note you will read

What was found, what you approved, what came down, what is still open. One page. A monthly version if you need something for a board.

A channel, not a ticket queue

A shared Slack channel. You ask about a finding, you get an answer, not a case number and a service level agreement.

From a call to your first findings, in about a week

A short call

What you sell, where you sell it, and who is allowed to sell it. Twenty minutes. If we are not the right fit we will say so on the call.

Connect your catalogue

Product photographs in, authorised sellers listed, price list attached. Upload it directly, or point us at your store and we will read it.

The first scan

Usually the same day. You see the entire run, including the pages we could not read, rather than a filtered highlight reel.

You start approving

Work the queue, tick what is real, decline what is not. Every decline makes the next scan sharper. We take it from there.

The questions people actually ask

Do you scrape Amazon and eBay?

No. They block crawlers deliberately and we are not going to work around that. We go through their rights-owner programmes instead, which is slower to set up and is also the only route that can get a listing removed. Anyone telling you they crawl those sites at scale is describing something that breaks the first time it matters.

What happens when you get one wrong?

You decline it and it never becomes a notice. Every finding shows its reasoning and, just as importantly, what it could not check. A false positive is a cost we would rather pay than miss a real copy, so the matcher leans towards showing you too much.

How long does a takedown take?

It depends entirely on who is hosting it. A Shopify store can go in days. A registrar in a jurisdiction that does not care can take months, or never happen. We tell you which one you are looking at when you approve it, not afterwards.

Who sees my catalogue?

Your photographs sit in private storage and are never used to train anything shared between brands. Links inside the dashboard are short-lived and signed. Your account is yours, and no other brand's data touches it.

Do I need a registered trademark?

Not to start scanning. You do need one for most marketplace brand registries, and for the strongest kind of takedown notice. If you do not have one yet, we will tell you exactly what it changes.

Find out what is already out there.

One scan of your catalogue, the whole result, and a walk through what we found and what we missed. If there is nothing worth acting on, we will tell you that too.

Book a scan Read the last full run

Private beta. We take a small number of brands at a time so every finding gets read.