Paste your text, add the pages you want to compare it with, and this plagiarism checker shows which passages appear word for word in each page, what share of your text they cover, and the longest copied run. It suits the two jobs SEOs and editors actually face: checking a freelancer’s article against the pages that already rank, and finding out whether a page you suspect copied you really did. It compares against the sources you give it rather than the whole web, and says so in every result.
On 12 October 2026 we tested it with a 144-word text that contained one sentence lifted from another text and one sentence quoted, in quotation marks, from Google’s robots.txt guide. With “ignore text in quotation marks” on, the quoted sentence dropped out of the count, and the checker reported 32 of the remaining 127 words, 25.2%, as a 32-word passage copied from the second text. Quoting with credit is not plagiarism, and the tool treats it accordingly.
How the plagiarism checker works
The checker compares sequences of words, not single words. It splits your text and each source into words, ignoring case, accents and punctuation, and looks for every run of seven consecutive words (you can choose five or ten) that appears in both. Overlapping runs join into passages, so a copied paragraph shows up as one passage, not twenty fragments.
- Sources. Up to five URLs, plus an optional second text pasted directly. For URLs, the doctor-seo.net server downloads each page and keeps its main text: the article or main content area when the page marks one, otherwise the body without navigation, header, footer and forms.
- Matching. Runs in your browser. Your own text is never sent to the server; only the URLs are.
- Quotes. Text inside quotation marks (“…”, “…”, «…», „…“) is left out when the option is on, so properly quoted material doesn’t count against you.
- Results. An overall percentage of your words found in at least one source, a row per source with its own percentage and longest passage, the longest passages themselves, and your full text with every match highlighted in that source’s colour.
How to read the result
The percentage is the share of your words that sit inside a matching run of at least seven words. Below 5% is usually set phrases, names and short quotes. Between 5% and 25% means some sentences match and are worth reading one by one. Above 25% means whole passages match. A number alone says nothing about intent: a product description that has to state the same specifications as the manufacturer will overlap, and so will a legal notice.
The run length matters more than the total. Ten scattered seven-word matches are often coincidence in a technical topic; one 60-word passage is not. That is why each source shows its longest passage, and why the passages are listed longest first.
What it finds and what it doesn’t
| Finds | Doesn’t find |
|---|---|
| Sentences and paragraphs copied word for word | Paraphrases and translations |
| Copies with changed capitalisation, accents or punctuation | Copies on pages you didn’t list |
| Partial copies inside longer texts | Text inside images or PDFs |
| Text on any public web page you list | Pages that block automated requests or need a login |
Web-wide plagiarism search needs a search engine’s index behind it. Google closed its Custom Search JSON API to new customers and will switch it off on 1 January 2027, so free tools that promise to search “the whole web” either pay for another index or search much less than they suggest. This checker is deliberately the other kind: exact, transparent, and limited to sources you choose. To find candidate pages, search Google for a distinctive sentence from your text in quotation marks, then paste the results here.
Duplicate content and SEO
When the same text exists on several URLs, Google’s documentation treats it first as a canonicalisation question: which version to show. Copying becomes a policy matter when a site is built on it: Google’s spam policies list scraped content, pages made from material taken from other sites without adding value, as a violation. In practice the risk of copied text is less a penalty than invisibility: when two pages say the same thing, Google tends to show one of them, and it may not be yours.
The short version
- The checker compares your text with up to five URLs and a second text, word for word, seven words at a time.
- It finds exact copies, including partial ones; it does not find paraphrases.
- Quoted text can be excluded, because quoting with credit is not plagiarism.
- Look at the longest passage, not only the percentage.
- Your text is compared in your browser and never stored.
FAQ
Is the plagiarism checker free?
Yes, with no account. Each check takes up to 5,000 words and five URLs, and there is a limit of 20 checks with URLs per connection every 10 minutes.
Is my text stored?
No. Your text never leaves your browser. The server only downloads the URLs you list, and keeps their text for 30 minutes so a repeated check is faster.
Why does it not search the whole internet?
Because that needs a commercial search index, and a free tool built on one either limits you to a handful of checks or searches far less than it implies. Comparing against pages you choose is exact and honest about its scope.
What does “count a match from 7 words” change?
It sets the shortest run that counts as a match. Five catches more, including common phrases; ten only flags clearly copied sentences. Seven is a sensible default for most articles.
Can it check text against a PDF or a Word file?
Not by URL. Copy the document’s text and paste it as the second text.
Where to go next
This tool is part of the Doctor SEO tools section. To review how a draft reads rather than where it came from, use the AI content checker. Duplicate content, canonical tags and how Google picks one version are covered in the technical SEO level of the free SEO course.
Sources
- Google Search Central, Spam policies for Google web search (scraped content), checked 12 October 2026.
- Google Search Central, How to specify a canonical URL with rel=”canonical” and other methods, checked 12 October 2026.
- Brave, Alternatives to the Google Search API before the 2027 shutdown, on the closure of Google’s Custom Search JSON API (vendor source), checked 12 October 2026.
- doctor-seo.net, first-party test of the checker on a 144-word test text, 12 October 2026.