No free detector reliably matches Turnitin, and chasing one is the wrong goal. Turnitin uses its own model, its own training data and its own thresholds, so another tool agreeing with it is coincidence rather than prediction. Running a second detector to guess the first tells you nothing a teacher can act on.
Key takeaways
- Turnitin’s detector is proprietary. No external tool has access to its model or thresholds, so none can be calibrated against it.
- Detectors frequently disagree on the same text, which is why one score cannot predict another.
- Students searching for a Turnitin match want reassurance before submitting. Teachers searching for one usually want corroboration. Neither gets it.
- Turnitin itself states its score is not proof and should not trigger an allegation on its own.
- For English language teaching the question matters less than it looks, because all perplexity-based detectors share the same bias against second-language writing.
Why no tool matches Turnitin
Turnitin’s AI detection is a closed system. The model, the training corpus and the score thresholds are not published, and Turnitin does not offer a public API for comparison testing. There is no way for another vendor to tune their output to match something they cannot see.
The deeper problem is that detectors genuinely disagree. The same essay can return a high score in one tool and a low one in another, because each weighs predictability differently and each was trained on different text. Peer-reviewed work in Patterns illustrates this directly: across seven detectors tested on the same 91 TOEFL essays, all seven agreed on only 18 of them, while 89 were flagged by at least one. That is a lot of disagreement on identical inputs.
So a free checker returning 80% tells you what that checker thinks. It does not forecast Turnitin.
What people are actually asking
This question comes from two directions and deserves a different answer each way.
Students usually want to pre-check work before submission, either to confirm honest writing will not be flagged or to edit until the score drops. The first is a reasonable anxiety, particularly for international students who have heard about false positives. The second is a losing game, because editing to beat one detector says nothing about another.
Teachers usually want a second opinion on a Turnitin score they are unsure about. That instinct is right, but a second detector is the wrong second opinion. Two unreliable probability scores do not combine into evidence. Drafts, version history and a conversation do.
What Turnitin says about its own score
Worth knowing before you treat the number as authoritative. Turnitin publishes a false-positive rate and advises that its AI score is an indicator rather than proof, and that it should not by itself trigger an academic misconduct allegation. The company positions the report as a prompt for a conversation with the student.
That is a more cautious position than many institutions take in practice, and it is useful to be able to point to when a score is being treated as a verdict.
The ESL problem that outranks the comparison question
If you teach English language learners, matching Turnitin is not the issue. The shared weakness is.
Every mainstream detector, Turnitin included, scores text partly on how predictable it is. Second-language writing is predictable by nature, because learners draw on a narrower vocabulary and practised structures. Research in Patterns found seven detectors wrongly flagged an average of 61.22% of human-written TOEFL essays, against 5.19% for native-speaker essays.
Finding a free tool that agrees with Turnitin would simply give you two tools making the same mistake about the same students. Our guide to why detectors flag non-native English speakers covers the mechanism.
What to do instead
If you are a teacher with a Turnitin score you distrust: ask for drafts, check version history, compare against the student’s earlier work, and have a short conversation about how they approached the task. That produces something you can actually stand behind.
If you are a student worried about a false positive: keep your drafts, work in a document with version history, and be ready to explain your choices. That evidence protects you far better than a reassuring score from a free checker.
For the full approach see our pillar guide to AI detectors for teachers and how teachers check for AI.
Read next
See the pillar guide to AI detectors for teachers, the honest comparison of AI detectors, and what AI detector teachers actually use.
Frequently asked questions
Which AI detector is closest to Turnitin?
None reliably. Turnitin uses a proprietary model with unpublished thresholds and no public API, so no external tool can be calibrated against it. Detectors frequently disagree on identical text, which is why one score cannot predict another.
Can I check my work on a free detector before submitting to Turnitin?
You can, but the result will not tell you what Turnitin will say. A low score elsewhere offers false reassurance and a high score causes needless worry. Keeping drafts and version history protects you far more effectively.
Is Turnitin more accurate than free detectors?
It is more consequential, because institutions act on it, but accuracy comparisons are hard to verify independently. Turnitin publishes a false-positive rate and states its score is not proof of misconduct, which is more transparency than most free tools offer.
Why do two detectors give different scores for the same essay?
Because each was trained on different text and weighs predictability differently. In one study of seven detectors run on the same 91 essays, all seven agreed on only 18, while 89 were flagged by at least one tool.
Does Turnitin treat ESL writing fairly?
It shares the same underlying weakness as other detectors. Perplexity-based detection penalises the predictable vocabulary and sentence patterns typical of second-language writing, so learners face a higher false-positive risk regardless of which tool is used.
Can a Turnitin AI score get a student into trouble on its own?
It should not. Turnitin advises that the score is an indicator rather than evidence and should not by itself trigger an allegation. Institutions are expected to investigate with process evidence and a conversation before taking any action.
Quick favour: how do you actually use AI in your teaching?
We are running the first real study of AI use among TEFL teachers. 12 questions, about 3 minutes, fully anonymous. Leave your email at the end and you will get the report before anyone else.
