The short version: the criterion that decides this is not accuracy, coverage or price. It is what the vendor does with the file you upload. A thesis chapter is unpublished work you have not yet deposited, and a checker that adds it to a comparison database can cause your own institution’s scan to match your thesis against a copy of your thesis. The table comes first; the reasoning comes after it.
| Tool | Can you use it yourself? | What it says about your upload | The catch |
|---|---|---|---|
| Turnitin Similarity | Normally only through an assignment your instructor sets up | Whether submissions are stored in a paper repository is an institutional setting | 🔴 Turnitin’s product and student-guide pages returned HTTP 403 to our requests in August 2026, so we cannot quote its current consumer position |
| Compilatio Studium | Yes, directly, with a free partial analysis | Documents “are not used as comparison material for other Compilatio users”, “cannot be viewed by teachers or other users” and “are not sold to third parties” | French company (Chavanod, Haute-Savoie), so your text is processed in the EU rather than in Canada |
| Grammarly Premium | Yes | Its plagiarism page describes coverage but not retention | Sits in a product family that also ships an AI Humanizer and an AI Detector |
| Quetext | Yes, free tier limited to 1,000 words | Its pricing page states coverage, not retention | The free plan bundles an AI Humanizer alongside the checker |
| Cite This For Me paper checker | Yes | 🔴 “The papers you upload will be added to our plagiarism database and will be used internally to improve plagiarism results.” | Disqualifying for an unpublished thesis chapter |
| Scribbr · Copyleaks | Yes, in principle | Not verifiable | Both sites returned HTTP 403 to our requests in August 2026; we will not describe terms we could not read |
Why retention is the criterion, and not accuracy
Every checker on that list will find an unattributed paragraph copied from a web page. They differ far more in what they keep than in what they find.
Consider the sequence. In March you paste chapter three into a free checker to reassure yourself. The vendor’s terms say uploads join its comparison database. In November your completed thesis is submitted through your programme’s system, and a scan reports a substantial match against a document in a third-party corpus. The matched document is your own chapter three. You now have to explain a match you created, using a service agreement you did not read, about a document nobody else has ever seen.
This is not hypothetical wording. Cite This For Me’s paper checker states in its own terms that “the papers you upload will be added to our plagiarism database and will be used internally to improve plagiarism results” — a clause we flagged in our review of APA citation generators for a Canadian thesis, where the same company appears for a different product.
And the opposite commitment exists in writing. Compilatio, whose teacher-facing product Magister is used inside institutions, states for its student product Studium that documents “cannot be viewed by teachers or other users”, that “they are not used as comparison material for other Compilatio users”, that “they are not sold to third parties”, and — the detail that matters most — that if your teacher uses Compilatio Magister, “they will not be able to find or access your Studium plagiarism analysis.” That is a vendor drawing an explicit wall between the student-side and the institution-side product.
The recommendation
Use a checker whose terms state that your upload does not become comparison material, and read that clause yourself before uploading anything. Of the tools whose terms could be read first-hand in August 2026, Compilatio Studium is the one that makes the commitment explicitly, and it offers a free first partial analysis showing a preview of the top three similar sources, which is enough to test the process on one section.
Runner-up for a quick check on a short piece: Quetext’s free tier, which covers up to 1,000 words with its DeepSearch checker — about three pages, so a section rather than a chapter. Read the caveat below about what else is in the box.
The zero-cost option that most students overlook: ask your programme whether it enables a draft or self-check submission. Many do, some do not, and the answer takes one email. It is also the only route where the report you see is the report your marker will see.
The prices, as far as they can be verified
Quetext publishes its list openly. As of August 2026 its site shows a free plan at $0 with the DeepSearch plagiarism checker limited to 1,000 words; a plagiarism-only plan at $9.99 per month for 50,000 words with downloadable originality reports; an Essential plan from $19.99 per month for 100,000 words; and a Professional plan from $29.98 per month for 200,000 words. Those are the figures shown in US dollars; the pricing page also offers a Canadian-dollar selector, so check the CAD figure before you assume a conversion.
For Turnitin, Scribbr and Copyleaks we are not going to quote prices. All three refused our requests, and a price copied from an aggregator is exactly the kind of stale figure that makes a comparison worthless six months later. If a number matters to your decision, get it from the vendor’s own checkout.
The uncomfortable detail about the free tiers
Quetext’s free plan includes the DeepSearch plagiarism checker, an AI content detector, an AI summarizer — and an AI Humanizer, at up to 2,000 words. Grammarly’s tool menu likewise lists a plagiarism checker, an AI detector and an AI humanizer side by side.
A humanizer exists to make machine-generated text read as though a person wrote it. Whatever you think of that as a product, it means the same company is selling you the check and selling somebody else the evasion. It is a portfolio conflict rather than a defect in the checker, but it is worth knowing which family of products your account sits inside — the same tension we set out in our ranking of AI thesis tools for Canadian students. It also matters for a different reason: a Canadian supervisory committee’s approval covers the tools you declared, and the answer to what AI use your committee will actually permit does not usually stretch to a rewriting engine.
What the coverage claims actually mean
Vendors describe their corpora in round numbers, and the round numbers are less comparable than they look.
- Grammarly states that its checker works “against billions of web pages” and against “ProQuest’s academic databases”, naming coverage of “millions of articles across 20+ popular research databases and websites, such as Wikipedia, Frontiers, PLOS One, ScienceDirect, SAGE Journals, PubMed, Elsevier, DOAJ, arXiv, and Springer.”
- Compilatio describes comparison against “billions of digital documents, among hundreds of billions of internet sites accessible online”, including private and open-access publications.
- Institutional tools add the one corpus no consumer product has: previously submitted student work at participating institutions. That is precisely why an institutional scan can find things a consumer check cannot, and why a clean consumer report is not a guarantee.
None of these numbers tells you whether the specific sources you used are indexed. A checker that has never seen your supervisor’s unpublished conference paper will return a clean result on a paragraph lifted from it.
Four questions to ask before you upload anything
- Does this service add my document to a comparison corpus? Find the sentence in the terms. If you cannot find it, treat the answer as yes.
- Does my programme store submissions in a paper repository? This is the setting that determines whether a document submitted twice can match itself. Your graduate coordinator knows; a vendor page does not.
- Where is my text processed? A thesis containing participant quotations is not just your writing. If your project went through research ethics review, your data-handling undertakings follow the text into any service you upload it to.
- What will I do with the number? If the honest answer is “feel better”, the check is not worth the risk. A similarity percentage is not a verdict, and no Canadian institution publishes a safe threshold — our guide to what similarity percentage is acceptable covers what the score does and does not measure.
Where the checker stops helping
Every tool on this page detects textual overlap. None of them detects the failure mode that actually ends degrees, which is patchwriting: a passage reworded closely enough to escape a match but not far enough to be your own argument, with the citation missing. A checker returns green and the examiner returns a misconduct referral.
The defence is not a better scanner. It is keeping quotations marked, sources attached and references consistent while you write, so that nothing enters the manuscript unattributed in the first place. Write your chapters in Tesify and keep citations and reference list consistent as the draft grows, while every argument remains 100% written by you.
Frequently asked questions
What is the best plagiarism checker for students in Canada?
The one whose terms state that your upload does not join a comparison corpus. Among the tools whose terms we could read in August 2026, Compilatio Studium states that explicitly. Your programme’s own draft-check option, where it exists, is better still.
Can I buy my own Turnitin account?
Turnitin is licensed to institutions and access normally runs through an assignment your instructor creates. Its product and student-guide pages returned HTTP 403 to our requests in August 2026, so confirm with your graduate coordinator rather than with a summary.
Turnitin or Grammarly for plagiarism detection?
They are not substitutes. Turnitin adds a corpus of previously submitted student work that no consumer product has; Grammarly is a consumer tool you can run yourself against web pages and ProQuest databases. A clean Grammarly report does not predict an institutional result.
Is a free plagiarism checker safe to use on my thesis?
Only if its terms say your upload is not retained as comparison material. Cite This For Me’s paper checker states the opposite in writing.
How many words can I check for free?
Quetext’s free plan covers up to 1,000 words with its DeepSearch checker. Compilatio Studium offers a free first partial analysis with a preview of the top three similar sources.
Why do these tools also sell an AI humanizer?
Because the consumer writing-tools market bundles them. Quetext’s free plan and Grammarly’s tool menu both list a humanizer next to the checker. Knowing that is part of choosing where your account lives.
Will a checker tell me whether my citations are correct?
No. It finds textual overlap. Whether an entry is formatted correctly, and whether a Canadian government or Indigenous source is even citable by a generator, is a separate problem.
Does using a checker count as academic misconduct?
Checking your own work is not misconduct. Uploading it to a service that keeps it can create a problem later, and using a rewriting tool to change what a scan finds is a different matter entirely.
Where is my text processed?
Compilatio is a French company based in Chavanod, in Haute-Savoie, so its processing sits in the EU. Grammarly’s own security documentation places its hosting in Amazon Web Services data centres in the US East region. Neither is Canadian.
Can my supervisor see that I ran a check?
Compilatio states that a teacher using its Magister product cannot find or access a Studium analysis. Other vendors make no equivalent statement, so do not assume the separation exists.
What score should I be aiming for?
There is no target. No Canadian university publishes an acceptable percentage, and a low score with one uncited paragraph is worse than a high score made entirely of quotations and a reference list.
What happened to Ouriginal?
Its service ended on 30 June 2026 and its domain now redirects to Turnitin. Any course outline or service page still naming it is out of date.