Direct answer
Compilatio is a plagiarism- and AI-detection platform sold to universities and schools, not a free consumer checker. Its report combines a similarity index (the share of your text that matches sources in its databases) with an AI-generated content score that estimates how likely passages are to be machine-written [1]. Neither number is a verdict: both are signals your instructor interprets against highlighted passages and cited sources, which is why a high similarity rate can appear on a fully honest paper and a low one can still hide a problem [2].
What Is Compilatio and What Does Its Similarity and AI Detection Actually Measure?
Compilatio began as a European (French) academic-integrity company and now bundles two distinct engines into one submission workflow: a document-comparison engine that matches your file against web pages, academic repositories, and its own archive of previously submitted student papers, and an AI-writing classifier trained to recognise statistical patterns typical of large language models [1]. The two outputs are reported separately because they answer different questions — "does this text already exist somewhere?" versus "does this text look machine-produced?" — and conflating them is the single most common misreading of a Compilatio report [1].
What the similarity engine returns is not a plagiarism score but an overlap measurement. Every matched string is highlighted and attributed to a source, so the same 24% can mean "four correctly quoted and referenced passages" or "one unattributed paragraph lifted from a repository," depending entirely on what the highlights contain [2]. Compilatio's own guidance stresses that the percentage is a starting point for a human reader, and that institutional thresholds (often set well above 10–15%) exist to trigger review rather than to condemn [2].
The AI component is probabilistic in a different way. It does not search for a source; it scores stylistic regularity, perplexity, and sentence-level predictability, then aggregates those into a document-level estimate with flagged segments [1]. Because academic writing is deliberately regular — standard phrasing, passive constructions, formulaic transitions — the classifier's confidence matters as much as its headline number [3].
How Should You Read a Compilatio Similarity Index and AI Score?
Read the similarity index source by source, not as one number. A Compilatio report lists each matching source with the percentage of your document it accounts for, so the productive question is whether that source appears in your reference list and whether the matched text sits inside quotation marks or a paraphrase you actually attributed [2]. A single source contributing 18% with no citation is a genuine problem; five sources contributing 3–5% each, all cited, is normal scholarly practice and reads as such to an experienced marker [2].
Read the AI score as a segment map rather than a verdict. Compilatio flags the passages driving the estimate, and those flags are frequently concentrated in exactly the places where human writing looks machine-like: literature-review summaries, methodology boilerplate, and definitional paragraphs assembled from standard textbook phrasing [3]. Non-native English writers are disproportionately affected, because the "safe" academic register they are taught to imitate overlaps heavily with what classifiers treat as machine-typical [3].
Practical reading order that experienced reviewers use: check the highest single-source contribution first, then the uncited highlights, then the AI-flagged segments, and only then the headline percentages [2][3]. If the AI flags cluster in passages you wrote yourself, the fix is usually stylistic specificity — concrete data, named examples, first-person methodological detail — not wholesale rewriting [3].
How Does Compilatio Compare With Turnitin for Checking Your Own Draft Before Submission?
The decisive difference is who can run the check. Compilatio and Turnitin are both institution-licensed systems: the account belongs to the university, submissions land in the institution's workspace, and results are surfaced to the instructor [1][4]. Turnitin's AI writing indicator, for instance, is presented to instructors inside the institutional workflow as a segment-level probability, and it displays *% instead of a precise figure when its confidence falls below the 20% threshold — a deliberate signal that low-confidence results should not be treated as findings [4]. Students cannot generate that report on their own draft through the official system [4].
The two engines also differ in database emphasis and in how much of the report is explained to the student. Compilatio's strength is European repository coverage and a report layout built around source-by-source attribution [1]; Turnitin's is breadth of web and publisher indexing plus a similarity report that instructors in the US, UK, Canada, Australia, New Zealand, and Ireland overwhelmingly recognise [4]. Neither vendor publishes a threshold that guarantees a "pass," because the threshold is set by the receiving institution, not the software [1][4].
For a student, the practical consequence is the same on both platforms: you cannot see your own report until after you have submitted, which is the worst possible time to discover a flagged methodology section [4]. That gap — between wanting certainty and having no official self-check — is exactly where an independent pre-submission preview becomes useful, and it is worth being explicit that such a preview is a diagnostic, not a substitute for your institution's own submission [1][4].
If the analysis above left you wanting to see your own document the way a marker will see it — similarity highlights and AI-flagged segments, side by side, before the deadline rather than after — turnitin0 exists for precisely that moment. Turnitin0 is an independent service, not affiliated with Turnitin, LLC, and it returns the same two PDFs your professor's system produces: a Turnitin AI detection report and a similarity report, delivered in under 15 minutes in 98% of cases, with no repository retention and no subscription.
※ Turnitin0.com - Actual Turnitin AI Report Cover, Score, Flag And Similarity Summary
FAQ
Is Compilatio the same as Turnitin?
No. They are separate commercial systems with separate databases and separate AI classifiers, and a university typically licenses one or the other [1][4]. A document that scores low on Compilatio can still be flagged on Turnitin, because the comparison corpora and thresholds differ.
What is a bad similarity percentage on a Compilatio report?
There is no universal number — institutions set their own review thresholds, and Compilatio itself frames the rate as an indicator requiring human interpretation rather than a pass/fail line [2]. Judge the highlights and their citations, not the headline figure.
Can I check my own draft in Compilatio before submitting?
Only if your institution gives you a student account with that permission; the platform is licensed to the university, so the report normally reaches your instructor first [1]. Independent preview services exist for students who need certainty before the deadline.
Does a high Compilatio AI score prove I used ChatGPT?
No. It is a probability estimate from stylistic features, and it can misfire on formulaic academic prose and on writing by non-native English speakers [3]. Treat flagged segments as a prompt to add specific, verifiable detail rather than as proof of misconduct.
Is Compilatio used outside Europe?
It is most established in European and francophone institutions, while Turnitin dominates in the US, UK, Canada, Australia, New Zealand, and Ireland [1][4]. Always confirm which system your own department actually uses before drawing conclusions from either report.