Reviewed and updated August 2026 against each tool’s own live product pages, checked 3 August 2026. Prices change frequently — confirm before purchasing.
Two things have changed enough since most “best plagiarism checker” lists were written that they are worth stating up front. Every serious tool now bundles AI-writing detection alongside text matching, and the vendors have consolidated to a degree that affects which “independent” options actually are independent.
Correction: Unicheck has shut down
An earlier version of this article named Unicheck as its favourite plagiarism checker. That recommendation is withdrawn, and the entry has been removed rather than quietly edited.
Unicheck’s service ended on 1 January 2025. The domain now redirects to Turnitin, which acquired it, and the page states plainly: “Unicheck service ended on January 1, 2025” while pointing former customers to Turnitin Similarity as the replacement. Turnitin notes that Similarity offers many of the same features, including similarity reports with percentage matches and source lists, and integration with Canvas, Moodle, Blackboard and Schoology.
If you are following an older guide that still recommends it, that is the reason the links do not work. It is also a reminder that this category consolidates quickly, and a list written two years ago is not a current list.
How to choose
The decision is mostly settled by two questions before you compare features at all.
Do you have institutional access? If your school licenses Turnitin, SafeAssign or similar through the LMS, use it. It is already integrated with your gradebook, it checks against your institution’s own archive of previously submitted papers — which catches more real reuse than any web-only checker — and results sit in the workflow your colleagues already use.
Are you checking your own students’ work, or your own writing? Tools built for individuals return a report to you. Tools built for institutions manage submissions, retain papers, and handle the administrative side. Buying the wrong category is the most common mistake here.
Then the practical criteria: what it actually indexes, whether student submissions are retained, whether it handles your file formats, whether AI detection is included, how the report presents matches, and what it costs per use rather than per month.
The one thing every good checker will tell you
Worth internalising before you read a single score: these tools do not detect plagiarism. They detect matching text.
Turnitin says so in its own words: “Turnitin does not check for plagiarism in writing. We do check all submissions against our database,” and the score is “simply the percentage of text… that matches other sources,” to be used “as a tool within your review process.” Blackboard describes SafeAssign as identifying “areas of overlap” and instructs that “instructors and students need to determine if the matching text is properly referenced.”
A correctly quoted and cited passage matches. A standard methods description matches. A reference list matches. None of that is misconduct — it is a functioning matcher doing its job. The number is where the review starts, not where it ends.
Paid plagiarism checkers
Turnitin — the institutional standard
Still the default in higher education and increasingly in secondary schools, and the one most LMS integrations are built around. Its real advantage is the database: billions of web pages, a large repository of previously submitted student papers, and subscription journals and publications.
Turnitin now also ships AI writing detection, and it goes further than most — its checker targets “an AI writing tool such as ChatGPT, text spinners, or AI bypassers,” and reports likely AI-paraphrased text as a separate category from likely AI-generated text. It suppresses scores in the 1–19% band and publishes a target false-positive rate under 1% for documents over 20% AI writing.
Best for: institutions. Limitation: it is not sold to individual teachers — access comes through your school, which is exactly why the rest of this list exists.
Quetext — the strongest individual paid option
Quetext is built for individuals and is unusually direct about the capability that matters most now: its checker “goes deeper, detecting paraphrased content,” using what it calls DeepSearch to scan “billions of sources across web pages, journals & academic publications.”
You can check 1,000 words free with no sign-up, which is enough to evaluate it properly before paying. Paid plans currently run from around $5.49 to roughly $24 per month depending on tier and billing term. It reports 10 million-plus users, 17 supported languages and sub-10-second scans, and integrates with Google Docs and via API.
Best for: a teacher buying their own tool. Limitation: no institutional submission management or student-paper archive.
Scribbr — thorough reports, priced per document
Scribbr markets itself as using “similar software as leading universities and publishers,” and bundles free access to its AI Detector and AI Proofreader with a premium check. Pricing is per-document rather than subscription, currently around $19.95, $29.95 and $39.95 by document length.
Disclosure worth making: Scribbr is a Learneo company — the same parent that owns Course Hero, QuillBot, Symbolab and Cram.com. That is not a reason to avoid it, but it is worth knowing that the company selling this plagiarism checker also sells one of the most widely used paraphrasing tools.
Best for: occasional checks on longer documents. Limitation: per-document pricing gets expensive across a whole class.
Grammarly — plagiarism checking inside a writing tool
Grammarly’s plagiarism checker is bundled into the wider writing platform rather than sold as a standalone product, and it now describes itself as AI-powered, detecting plagiarism alongside grammar, spelling, punctuation, conciseness, readability and word choice in a single scan.
Best for: teachers who already use Grammarly and want originality checking as part of existing feedback. Limitation: the index is not built for academic submissions the way Turnitin’s is, and it has no student-paper archive — so treat it as a first-pass screen rather than an institutional check.
Free plagiarism checkers
Free tools are genuinely useful for a quick check on a single suspicious passage. They are not substitutes for an institutional tool, for one structural reason: they cannot see your school’s archive of previously submitted papers, which is where student-to-student reuse actually shows up.
Duplichecker — the most capable free option
Free checking up to 10,000 characters per scan, accepting pasted text or .docx, .txt and .pdf uploads, with support across a long list of languages including Spanish, Portuguese, German, French, Arabic and Japanese.
It has expanded well beyond simple matching: current scan settings include AI Detection, options to exclude quotes and bibliography, and Deep Scan and Academic Scan modes — though several of those are marked PRO rather than free.
Best for: a fast free check with useful exclusion controls. Limitation: the character cap means longer essays need splitting, and the best modes are paid.
SmallSEOTools — simple and unlimited enough for spot checks
A long-running free checker that compares text “against billions of online sources” and returns a report with matched sources. It accepts direct paste or uploads from local and cloud storage in .doc, .docx, .txt, .odt and .pdf, and now also offers a Deep Plagiarism Checker, a grammar checker and AI content detection.
Best for: quick spot checks with no account. Limitation: built for web content rather than academic work, with no academic database, no archive and an ad-supported interface.
At a glance
| Tool | Cost | AI detection | Best for |
|---|---|---|---|
| Turnitin | Institutional licence | Yes — incl. paraphrase & bypass | Schools and universities |
| Quetext | Free 1,000 words; ~$5.49–$24/mo | Yes | Individual teachers buying their own tool |
| Scribbr | ~$19.95–$39.95 per document | Yes | Occasional checks on long documents |
| Grammarly | Bundled with Grammarly | Yes | Existing Grammarly users, first-pass screening |
| Duplichecker | Free (PRO tiers paid) | Yes | Free checks with exclusion controls |
| SmallSEOTools | Free | Yes | Fast spot checks, no account |
A note on who owns what
This category has consolidated to a degree that is worth knowing when you read any recommendation list, including this one.
- Turnitin acquired Unicheck and shut the service down on 1 January 2025
- Learneo owns Scribbr, QuillBot, Course Hero, Symbolab and Cram.com — meaning the same parent sells a plagiarism checker and a leading paraphrasing tool
- Chegg owns Mathway, and bundles plagiarism detection into its Study Pack
None of that makes any individual tool worse. It does mean that a list of “independent alternatives” is often shorter than it appears, and that a recommendation written a couple of years ago may be describing a company that no longer exists as such.
The types of plagiarism these tools catch
Worth knowing because the tools handle them very differently:
| Type | What it is | Caught by matching? |
|---|---|---|
| Direct | Copied word-for-word from a source without attribution | Yes — this is the easiest case |
| Mosaic | Phrases woven in from several sources with light edits | Often, though matches appear as scattered fragments |
| Self-plagiarism | Reusing your own previously submitted work | Only where an archive holds the earlier submission |
| Accidental | Poor paraphrasing or missed citation, without intent | Yes — and this is why a match is not a verdict |
| Paraphrasing | Reworded source material, still uncited | Weakly by matching; increasingly by AI-paraphrase detection |
| Contract cheating | Work produced by someone else | No — matching cannot catch original text written to order |
That last row is the important one, and it is why detection alone is never a complete answer. Text that was written from scratch by a third party matches nothing, because it is genuinely original — just not the student’s. The defences against it are assessment design and knowing your students’ work, not software.
An earlier version of this page carried an outbound link to an essay-writing service. That link has been removed. It should never have been on a page about academic integrity.
What to do when a checker flags something
The report is the beginning of a process, not the end of one, and how you handle the next step determines whether the tool improves anything.
Read the matches before forming a view. Open the highlighted passages and look at what they are. A block of matched text that turns out to be a correctly formatted quotation with a citation is not a finding. A matched reference list is not a finding. Scattered three-word matches on common academic phrasing are not a finding either.
Work out which kind of problem you are looking at. A single uncited paragraph lifted wholesale is a different conversation from a paper stitched together from six sources, which is different again from a student who paraphrased badly because nobody taught them how to paraphrase well. The first is misconduct; the last is a teaching opportunity that a misconduct process handles badly.
Ask the student. Every vendor in this category says the same thing about its own output, and they say it because it is true: the tool provides data for a human to interpret. A conversation resolves a surprising number of these — students can usually produce their sources, their notes and their drafting history if the work is theirs.
Follow your institution’s process for anything serious. Formal procedures exist to protect students and staff alike, and handling a genuine case informally to be kind tends to be unfair to everyone else.
Use it preventatively where you can. Where a tool allows students to see their own report before the deadline, letting them do so changes its role entirely — Blackboard notes that instructors decide whether students can see SafeAssign results, and describes SafeAssign as working “as both a deterrent and an educational tool.” A student who fixes their own citation problem before submitting is the best outcome available, and it costs you nothing.
Using the results fairly
- Open the report, never act on the number. Composition is everything — check what is actually highlighted
- Exclude quotes and bibliography where the tool allows it, before reading the score
- Expect high scores on some legitimate work — technical methods sections and heavily-cited literature reviews match a lot by nature
- Treat AI-detection flags with more caution than match reports — vendors publish false-positive rates for a reason, and formulaic writing triggers them
- Ask the student before concluding. Every vendor in this category says the same thing: the tool informs a human decision, it does not make one
- Tell students what you use and let them see reports where possible — it works better as a deterrent and a teaching tool than as a trap
What makes a good plagiarism checker
If you are evaluating something not on this list, these are the criteria that actually separate a useful tool from a box that returns a percentage.
The size and type of index. This is most of the product. A checker that only crawls the open web will miss subscription journals, and every web-only checker misses the single richest source of student-to-student reuse: your own institution’s archive of previously submitted papers. That archive is the reason institutional tools outperform free ones on the cases you most want to catch.
How the report presents matches. A percentage on its own is close to useless. You want each match highlighted, numbered against a listed source, and openable so you can see the original. Controls to exclude quotations and bibliographies before scoring matter enormously, because without them a well-referenced essay looks alarming.
What it does with submitted work. Some tools retain submissions to strengthen future checks; others do not. Both are defensible, but you should know which you are using — retention has privacy implications worth being transparent with students about, and non-retention means resubmitted work will not be caught.
File handling. Support for the formats you actually receive, and enough capacity to take a full essay without splitting it. Note that no text matcher can read an image of text, so a scanned PDF without a text layer returns nothing regardless of the tool.
Whether AI detection is included, and how honestly it is described. A vendor that publishes its false-positive rate and suppresses low-confidence scores is being more useful than one advertising perfect accuracy. Turnitin’s decision not to attribute any percentage below 20% is a good example of the former.
Cost per actual use. Per-document pricing is fine for occasional checks and expensive across a class of thirty; a monthly subscription is the reverse. Work out roughly how many documents you will check in a term before comparing headline prices.
Frequently asked questions
What is the best free plagiarism checker for teachers?
Duplichecker is the most capable of the free options, offering scans up to 10,000 characters with quote and bibliography exclusion and AI detection, though its deeper scan modes are paid. SmallSEOTools is the simplest for a fast check with no account. Neither can see your institution’s archive of previously submitted papers, which is the biggest gap versus an institutional tool.
Is Unicheck still available?
No. Unicheck’s service ended on 1 January 2025. The domain now redirects to Turnitin, which recommends Turnitin Similarity as the replacement and notes it integrates with Canvas, Moodle, Blackboard and Schoology.
Can teachers use Turnitin without their school?
Not usually — Turnitin is sold as an institutional product through schools and universities rather than to individual teachers. Individuals generally use tools like Quetext, Scribbr or Grammarly instead.
Do plagiarism checkers detect AI writing?
Most now include some form of AI detection, and Turnitin’s goes furthest by separately reporting likely AI-paraphrased and bypassed text. Accuracy varies and false positives are real — Turnitin publishes a target under 1% for documents over 20% AI writing and suppresses scores below 20% entirely. Treat these flags as a prompt for a conversation, not as proof.
What is a bad plagiarism score?
There is no universal number, and Turnitin deliberately publishes no threshold at all. A score built from cited quotations and a reference list means something completely different from the same figure drawn from one uncited source. Read the report.
Students are on the other side of the same question. Our guide to paraphrasing tools and academic integrity sets out what those tools do and where their use crosses into plagiarism.
Josh Hutcheson — Editor, PriorityLearn
Josh researches, writes, and updates the answers on PriorityLearn, checking each one against current tools, official sources, and real school policies — and flagging what varies by state or district. About PriorityLearn →