Website Similarity: compare your site against sites you choose

Compare your site against sites you choose: copied wording, copied design, same template. Give it your website and the addresses you want it compared with. It opens each one, scores how alike it is, and tells you why: copied wording, page layout, look, custom code, shared image files, and whether the match is just the same off-the-shelf template. Each result comes with plain-English reasons, your copied sentences, a side-by-side picture and a PDF you can keep. No login. It reads only the sites you list, and only the pages their robots.txt allows.

Quick start

Add your site under Your websites and the sites to compare under Compare against these sites, then press Start.

{
  "startUrls": ["https://www.yourcompany.com"],
  "compareWith": ["https://competitor-one.example", "https://competitor-two.example"],
  "checks": ["content", "design"]
}

compareWith is required: the Actor does not search for sites, it compares your site with the ones you name. Every site you list is scored and returned, even when the score is low.

What you get

One row per site you listed, best match first:

Field	Meaning
similarityScore	0 to 100. 70 and up is a clear copy. 40 to 70 is a look-alike or a shared template
verdict	copied-content, copied-design, same-template, lookalike
reasons	Plain English, strongest first: "7 of your 10 sentences word for word", "Same page layout (88%)", "Uses 3 of your image files", "Both use the Astra theme: likely a template, not a copy"
copiedSentences	Up to 5 of your own sentences found word for word on their site
contentOverlap, topicSimilarity	Share of your 5-word phrases on their pages, and how alike the subject matter is
layoutSimilarity, visualSimilarity, paletteSimilarity	Where the blocks sit, how the first three screens look, how alike the colours are
sharedFiles, sharedClasses, sharedFonts	Identical image or style files (including your files loaded straight from your server), custom CSS names both pages use, shared fonts
templates	Theme or site builder detected on each site
archive	When the Internet Archive first saved each page, and which was earlier (copy verdicts)
comparisonImageUrl, reportUrl	A side-by-side picture and a PDF evidence report, for scores of 60 or more
foundVia	Always your-list (the sites you named)

The run summary (key SUMMARY) lists websites checked, sites compared, what could not be read, and the cost.

How it decides

• Copied content. It takes distinctive sentences from your pages, reads the pages of each site you listed, and counts your sentences and 5-word phrases on them.
• Copied design. It opens both pages in a browser and compares what each block on the first three screens is (heading, text, image, button, card) and where it sits, the look of the screen, the colours, the custom CSS class names, and byte-for-byte identical image and style files. Class names that many unrelated sites use (page builders, plugins, frameworks) are ignored, so two sites on the same builder are not called copies of each other.
• Same template. Two sites on the same theme (for example the same WordPress or Shopify theme) look alike by design. These are labelled same-template and are not called copies unless they also share files or custom code.
• Look-alikes. Sites with a close layout or the same subject matter.

On our own test set of unrelated small-business sites, 99% of pairs score below 25. Plain clones, new-wording copies of a design, and re-skinned copies of the wording all score 70 or more. The score is an estimate: read the reasons and the picture before you act.

Pricing

Pay per event, no subscription:

• Website checked: $0.12 per website you ask about (cheaper on higher Apify plans). It covers reading your site and comparing it with the sites you list. A website that cannot be read is free.
• Similar site: $0.02 per compared site returned. Repeats in monitor mode and unreadable sites are free.

A typical check against 2 listed sites costs about $0.16. The run summary shows what the run cost us to execute, for transparency.

Watch for new copies

Turn on Only new matches since the last run and schedule the Actor weekly. It remembers what it already told you (in a named store in your account) and returns only new matches or ones whose score rose by 15 points or more. Rows carry change: "new" or "rising".

Options

Input	What it does
startUrls	Up to 50 of your websites to check (the older name websites still works)
compareWith	Up to 100 sites to compare them with. Required
checks	content, design, or both. Design opens each page in a browser, so it takes longer
minScore	Used for the summary report's look-alike count. Every site you list is returned with its score
pagesPerSite	Pages read on your site (home plus linked pages), 1 to 10
onlyNew, watchName	Monitor mode and the name of its memory
saveImages, makeReports	Side-by-side pictures and PDF reports

Inputs from earlier versions (findSimilar, maxCandidates, country) are accepted and ignored. If findSimilar is on, the run says so in its status message.

Limits, stated plainly

• It does not search the web. It compares your site with the sites you name. To find who copied you, you need the addresses first.
• Sites that block automated visits (bot checks, a refusal such as HTTP 403 or 429) are reported as unreadable and left alone. The Actor does not retry them from another address and does not use a proxy. Unreadable sites you listed are returned without a charge.
• robots.txt is respected. Pages a site does not allow are not read. Requests to one site are made one after another.
• Free plan: up to 3 websites per run and 5 compared sites per website. Any paid Apify plan removes both limits.
• Scores for look-alikes are less certain than for copies. Use them to find competitors, not as proof.

Responsible use

Use it on business websites. It reads public pages only, never logs in, and never tries to find out who owns a site. Results are automated estimates, not legal advice. If you plan to send a takedown request, keep the PDF report and check the evidence yourself first.