Is this a Copyscape replacement?

No. We only compare the two inputs you give. There is no web-wide plagiarism database.

Is navigation counted?

On URLs, nav, footer, and similar chrome are stripped. We compare main or article text. Paste mode does not strip chrome for you.

What does vocabulary only mean?

The two texts share words but no five-word phrase. The percent stays at 25% or below.

When is it Near-duplicate?

Similarity 90% or higher, which requires shared five-word passages.

Can I use this as legal proof?

No. It is a writing aid, not a forensic tool.

Why is the score 0 on different topics?

No five-word phrase was shared, so the run is vocabulary only and the percent stays at 25 or below. Unrelated sentences about orchards and qubits returned 0. A 90 needs shared passages, not a similar subject.

What does Comparing on the button mean?

Both sides are still being fetched or scored. Page URLs loads two documents and strips chrome. Paste text compares the two fields as given, each capped at 100,000 characters. The percent appears when both sides finish.

Why is a URL pair thin when I can see a long page?

Main text after nav, footer, and related-post strip was under 50 words on that side. A shell or a JS article looks thin. Paste the article body, or use a URL that already sends the copy in HTML.

Duplicate Content

Loading tool…

About Duplicate Content

Choose Page URLs or Paste text and click Compare. URL mode extracts main or article text and strips nav, footer, share widgets, and related-post chrome. If a fetched page has under 50 words of main text, you get a thin-content notice. Paste mode compares the two blocks as given (100,000 character cap per field).

This is pairwise similarity, not Copyscape and not a web-wide index. High scores need shared five-word passages, not only similar topics. Overlapping passages are highlighted (up to 12). Verdicts: Near-duplicate, Substantial shared passages, Some shared passages, or Little shared language.

When to use it

Use it after you clone a landing page or spin a location variant, when two URLs should differ but still share the same five-word blocks. That is also the check before you publish a rewrite that only swapped nouns.

Use Paste text for drafts that are not live. Use Page URLs when both documents already render HTML. It is not Copyscape and it does not search the rest of the web.

How to use Duplicate Content

1Enter two URLs or paste two texts. Use Page URLs for live pages, Paste text for drafts.
2Click Compare. The button shows Comparing until both sides are scored.
3Read the percent and verdict. Highlights mark the overlapping passages.
4Rewrite the shared blocks. Then Compare again.

How similarity is counted

Text is split into five-word phrases. If any phrases are shared, the percent is mostly phrase coverage (85%) plus a little vocabulary overlap (15%). If no five-word phrase is shared, the score is capped at 25% and labeled as topic vocabulary only.

Different languages will look dissimilar even when the meaning matches. This is not a forensic or legal report. Rewrite the highlighted blocks if you need the two pages to diverge.

How to read the results

The percent is pairwise only. Verdicts are Near-duplicate (90% or higher), Substantial shared passages (50 to 89), Some shared passages (25 to 49), Little shared language (under 25 with shared five-word phrases), or No shared passages found when the score is vocabulary only.

Up to 12 overlapping passages are highlighted. URL mode notices Text A or Text B is only N words when that side has under 50 words after chrome strip. Two near-identical sitemap sentences scored 90 and Near-duplicate. Unrelated apple and qubit sentences scored 0 and vocabulary only.

Why this matters

Engines that see the same five-word passages on two URLs have no reason to keep both as distinct results. The second URL is the one that usually drops, even when titles differ. A topic match without those phrases is capped at 25% so a shared subject does not look like a clone.

Nav and footer chrome is stripped on URLs so a shared menu cannot manufacture a high score. Paste mode does not strip that for you, so two full HTML dumps will look more alike than two article bodies.

Limits and privacy

Page URLs fetches two public documents. Loopback and reserved IPs are rejected before either side loads. Paste is sent for that Compare only. About 20 compares per minute on your network. There is no plagiarism index and no login.

Common mistakes

Pasting whole page HTML, including nav, and reading 90% as article theft. Strip chrome or use Page URLs.

Comparing two languages and calling the tool broken. Meaning is not scored. Only English-like tokens and five-word phrases are.

Treating Little shared language as proof Google will keep both URLs. This pair never sees a third page.

FAQs

Is this a Copyscape replacement?: No. We only compare the two inputs you give. There is no web-wide plagiarism database.
Is navigation counted?: On URLs, nav, footer, and similar chrome are stripped. We compare main or article text. Paste mode does not strip chrome for you.
What does vocabulary only mean?: The two texts share words but no five-word phrase. The percent stays at 25% or below.
When is it Near-duplicate?: Similarity 90% or higher, which requires shared five-word passages.
Can I use this as legal proof?: No. It is a writing aid, not a forensic tool.
Why is the score 0 on different topics?: No five-word phrase was shared, so the run is vocabulary only and the percent stays at 25 or below. Unrelated sentences about orchards and qubits returned 0. A 90 needs shared passages, not a similar subject.
What does Comparing on the button mean?: Both sides are still being fetched or scored. Page URLs loads two documents and strips chrome. Paste text compares the two fields as given, each capped at 100,000 characters. The percent appears when both sides finish.
Why is a URL pair thin when I can see a long page?: Main text after nav, footer, and related-post strip was under 50 words on that side. A shell or a JS article looks thin. Paste the article body, or use a URL that already sends the copy in HTML.