Is your text already out there on the web?
We take the distinctive phrases out of your draft, search the live web for them, and show you the passages that match along with a link to each source.
Built to spot text from tools like these
- ChatGPT
Used by students at the world's top universities
- Harvard
- MIT
- Stanford
- Oxford
- Cambridge
- ETH Zürich
- Yale
- Princeton
- Imperial College London
- TUM
- Columbia
- UCL
- HKU
- Penn
- LSE
- HKUST
- Cornell
- Edinburgh
- CUHK
- Brown
- NUS
- Dartmouth
Based on individual use by students and graduates · no affiliation
What people say
4.538
Real reviews from people who use the service. Emails are partly hidden, and reviews are shown in the language they were written in.
I ran a dissertation chapter through before submission and it caught two paragraphs where I'd paraphrased too closely to the source. It linked the original, so fixing the citation took five minutes.
We QA every article before it reaches the client. The matched-passage view is what sells it — I don't have to read the whole report, just the flagged parts.
Used it to self-check an essay before turning it in. Found one quote I'd forgotten to attribute. Cheap insurance for something that could have gone very badly.
Journal submissions are stressful enough. Running the manuscript here first means I know exactly what a reviewer's similarity check will show. No surprises so far.
I run it on quote-heavy drafts. It separates quoted text from everything else.
I use it on student assignments. For me, it separates quoted text from everything else.
In my day-to-day work, it separates quoted text from everything else. I run it on quote-heavy drafts.
For me, I use it on student assignments. It shows the source links right there.
I check citations before submitting. For me, it shows the source links right there.
What this checker does
A check here is not a lookup in a stored database. Each run pulls the most distinctive phrases out of your text, searches the live web for them, and reports back the passages that turn up somewhere else, the page they appear on, and a similarity figure for the text as a whole.
That is also why the result is not a verdict. A high figure can simply mean you quoted a well-known source and quoted it properly; a low one does not clear a single unattributed sentence. Open each match, look at the page behind it, and decide what needs a citation.
Searchable public pages are the whole of the comparison set. Subscription journal archives and unpublished student papers are not in it, so before a thesis submission use your institution's official check as well.
Not just a number — where it matched.
Paste your draft
A report, article or thesis draft. The checker picks out the distinctive phrases to search for.
Search the web
Those phrases are searched on the live web, and public pages carrying the same or nearly the same sentences are collected.
Check sources and cite
Open each match and its link, and add a citation wherever one is needed.
Next step: sort out the sources.
Found a matching passage? Credit it first — in APA, MLA or Chicago.
If you need to put a passage in your own words rather than quote it, keep the meaning and change the expression.
Wondering whether it was written by AI, not copied? That's a different measure — use this one.
To avoid lifting whole passages from your reading, summarize it first and write from the summary.
Common questions
- How does the check work?
- It pulls the distinctive phrases out of your text, searches the live web for each of them, and lists the passages that match with a link to the page they came from. Nothing is compared against a stored archive of past submissions.
- Is there a free version?
- No. Every run performs a real web search that costs money, so each check spends one checking credit. If a check fails to return a result, the credit goes back to your account.
- How much similarity counts as plagiarism?
- There is no fixed line. A passage you quoted and cited properly will match, and so will a common turn of phrase. Read the matching passages, work out which ones need attribution, and cite them.
Before you submit, check what overlaps.
Each check uses one credit, and credits are one-off packs, not a subscription. If a check fails, the credit comes back.