How Moodle Plagiarism Plugins Compare: Turnitin, Urkund/Ouriginal, and Compilatio

Open your Moodle gradebook, click into an assignment, and you might see a little colored percentage sitting next to each submission. That number didn’t come from Moodle. It came from a plugin your school bolted on — usually Turnitin, Ouriginal, or Compilatio — and the three don’t work the same way under the hood.

If you’re an instructor choosing one, or a student trying to understand the report staring back at you, the differences matter. This is a plain comparison of how the big three integrate with Moodle, what their reports actually tell you, and where AI detection fits into the picture.

Key takeaways

• Moodle has no built-in originality checker. It exposes a Plagiarism API, and a third-party plugin does the real work.
• Turnitin has the deepest Moodle integration and the largest matching repository. Ouriginal (formerly Urkund, now owned by Turnitin) and Compilatio are strong alternatives, especially across Europe.
• A similarity score measures overlap with sources. An AI-writing score is a separate estimate. Different systems, different questions.
• All three produce color-coded reports, but they handle quotes, bibliographies, and translated text differently.
• The report is where a human conversation starts — not a verdict on its own.

First, what Moodle actually does

Moodle is a learning management system. It collects submissions, stores them, and runs grading workflows. What it does *not* do, on its own, is compare your essay against the rest of the internet or a database of past papers.

What it offers instead is a Plagiarism API — a standardized hook that lets external services plug into the submission flow. When a student uploads a file, Moodle hands it to whichever plugin the admin has enabled. The plugin sends it off, gets a result back, and Moodle displays that result inside the assignment view. So the “Moodle plagiarism score” is really a Turnitin score, an Ouriginal score, or a Compilatio score wearing a Moodle badge.

That distinction is the whole reason a Moodle plagiarism plugin comparison is worth doing. You’re not comparing Moodle features. You’re comparing three separate companies’ engines that happen to share the same doorway.

Turnitin

Turnitin is the one most North American students recognize on sight. Its Moodle integration (the Moodle Direct plugin, plus the newer LTI-based setup) is mature and tightly woven into the assignment dropbox.

Database. Turnitin’s biggest advantage is scale. It holds an enormous repository of previously submitted student papers, alongside web content and licensed academic journals. If a passage was lifted from a paper another student turned in three years ago at a different university, Turnitin has a real shot at flagging it.

The report. Submissions come back with a Similarity Report showing a percentage and color band, with matched text highlighted and linked to its apparent source. Instructors can filter out quotes and bibliographies, exclude small matches, and adjust what counts.

AI detection. Turnitin added a separate AI writing indicator that lives in its own tab — not folded into the similarity number. We’ve covered what that score means in more detail elsewhere on the blog; the short version is it estimates the percentage of text it believes was AI-generated, and it’s a probability, not a confession.

If your institution is large and English-dominant, Turnitin is often the default for a reason. The cost and the data-retention model (papers added to its repository) are the trade-offs people debate.

Ouriginal (formerly Urkund)

Here’s where the naming gets messy. Urkund merged with PlagScan in 2020 to become Ouriginal, and in 2021 Turnitin acquired Ouriginal outright — so you may see Urkund, Ouriginal, or even Turnitin branding depending on how old your school’s documentation is. For Moodle purposes, treat “Urkund” and “Ouriginal” as the same lineage.

Database. Ouriginal grew up in Europe and reflects it. Its matching draws on a broad multilingual corpus, published works, and a shared archive of submitted documents. For institutions teaching in Swedish, German, French, or other non-English languages, that coverage is a genuine differentiator.

The report. The analysis report flags matched sections with source links and a significance rating, and it’s built to surface *meaningful* overlap rather than every trivial three-word coincidence. Reviewers tend to like that it leans toward fewer, more relevant matches.

Integration. The Moodle plugin connects submissions to an institutional Ouriginal account, often routing through an analysis email address or LTI connection depending on the version. It’s less ubiquitous in the US than Turnitin but well-established in European higher ed.

Compilatio

Compilatio is French in origin and dominant in francophone institutions, though its reach extends well beyond. Its Moodle plugin (Compilatio Magister / the LTI integration) is purpose-built for the LMS workflow.

Database. Compilatio matches against web sources, a document library, and an archive of institutional submissions. Like Ouriginal, it’s comfortable with multilingual content and markets cross-language detection as a feature.

The report. Compilatio’s interface leans visual and instructor-friendly: a similarity percentage, a breakdown of suspect passages, and tools to distinguish genuine plagiarism from quoted or referenced material. It also offers analytics aimed at helping teachers spot patterns across a class rather than judging one paper in isolation.

AI detection. Compilatio has shipped its own AI-content detection alongside similarity checking, again as a distinct signal. Same caveat applies as with the others: the number is an estimate.

Mini-scenario: same essay, three plugins

Picture a student named Dana who submits a 2,000-word history essay. It’s her own work, but she quotes two primary sources at length and includes a full bibliography.

• Turnitin might return 24% — until the instructor toggles “exclude quotes” and “exclude bibliography,” at which point it drops to 6%. The raw number looked alarming; the filtered one is unremarkable.
• Ouriginal might show a lower headline figure to begin with, because its scoring de-emphasizes common quoted phrasing, and the report points the instructor straight to the two passages worth a look.
• Compilatio might land somewhere in between, with a visual highlight reel that makes it easy to see the quotes are properly attributed.

Three plugins, three numbers, one honest essay. (Those figures are illustrative, not benchmark results — the point is the spread, not the digits.) That’s the lesson: the percentage is an input to human judgment, not a grade. If three systems can disagree about Dana’s quotes, none of them should be treated as the final word.

Similarity is not the same as “AI”

This trips up students and instructors constantly, so it’s worth stating flatly. A similarity score asks: *how much of this text matches something we’ve seen before?* An AI score asks: *how statistically likely is it that a machine wrote this?* They run on different logic.

AI detectors lean on signals like perplexity and burstiness — roughly, how predictable each word is and how much the sentence rhythm varies. Human writing tends to be lumpier and less predictable; a lot of unedited AI output is smooth and even. That’s a statistical guess about *style*, not a database match against *sources*. If you want the underlying math, our piece on perplexity and burstiness walks through it.

Two consequences follow. A paper can be 100% original and still draw an AI flag because its prose happens to be flat and even. And a paper can be lightly AI-assisted yet pass, because it reads with natural variation. Neither tool is reading minds.

It’s also worth remembering these systems aren’t infallible. OpenAI pulled its own AI-text classifier in July 2023, citing a low rate of accuracy. And a 2023 study from a Stanford team (Liang and colleagues, with James Zou as senior author, published in the journal *Patterns*) found that several detectors disproportionately flagged writing by non-native English speakers. Detectors output probabilities, full stop — useful as a prompt to look closer, not as proof.

How to choose (or how to read the verdict)

If you’re an admin comparing plugins, the practical questions are: What languages do your students write in? How big is your matching need — do you require the largest repository, or is meaningful-match precision more valuable? What’s the budget, and how do you feel about your students’ papers being added to a shared archive? Turnitin tends to win on raw database size and US familiarity; Ouriginal and Compilatio compete hard on multilingual coverage and report clarity.

If you’re a student looking at a flagged report, the move is the same regardless of which plugin produced it: open the report, read the highlighted passages, and figure out whether the matches are real overlap or just quotes and common phrasing. If it’s an AI flag, keep your drafts and notes — process is your best evidence. You can read more on detection and writing across the blog.

Frequently asked questions

Does Moodle check for plagiarism by default? No. Moodle has a Plagiarism API but does no checking itself. A school must install and pay for a plugin like Turnitin, Ouriginal, or Compilatio. Without one, files are stored but never compared.

What’s the difference between a similarity score and an AI score? Similarity measures overlap with existing sources. An AI score estimates the chance the text was machine-written, from statistical patterns. They’re separate systems answering separate questions — a paper can score low on one and high on the other.

Which plugin is most accurate? There’s no universal winner. Turnitin has the largest repository; Ouriginal and Compilatio excel with multilingual content. For AI detection, all produce probabilities, and accuracy shifts with document length and style.

Can these plugins detect paraphrased or translated text? Partly. Modern checkers do some paraphrase and cross-language matching, so word-swapping is less reliable than students think — and uncited ideas remain a citation problem even when the similarity number is low.

How should I respond if my Moodle report flags my work? Read the matched passages, not just the percentage. High scores are often quotes and references. For AI flags, gather drafts and version history. The report starts a conversation; it doesn’t end one.

Closing

Whichever plugin sits behind your Moodle dropbox, the smartest thing you can do is understand what its number means — and what it doesn’t. Originality is a habit of citing well and writing in your own voice, not a percentage to game. If you want to understand how AI-detection signals read your draft so you can write more naturally and clearly, try PaperBleach.