3 October 2026 · 7 min read
It is one of the most common stories we hear from students in Nepal. They ran their assignment or thesis chapter through a free plagiarism checker, got a comfortably low number, submitted with confidence — and the department's report came back several times higher. The reverse happens too: a free tool returns an alarming score on work that turns out to be fine. This post explains why free checkers and Turnitin disagree, when a free check is perfectly adequate, and when it is a false comfort.
Every plagiarism checker works the same way at heart: it breaks your document into runs of words and looks for matching runs in a collection of other text. The score is the share of your document that found a match. Which means the single biggest factor in the number is not your writing — it is the size and content of the collection being searched.
Free checkers generally search the open web: pages a search engine can reach. That is a large collection, but it is missing exactly the material that matters most for academic work.
In Nepal this gap is especially sharp. A large share of the copying that actually gets students into trouble is from earlier theses and project reports in the same department — reused literature reviews, borrowed organisation profiles, inherited methodology sections. That material is rarely on the open web, so a free checker reports nothing, and the department's report lights up.
Even where two tools search similar text, they do not count matches the same way. Some count every three or four matching words; others ignore short runs. Some exclude quotations and bibliographies automatically; others count them. Some treat lightly reworded text as a match; others need an exact run.
Turnitin's report is also configured by whoever runs it — exclusions for the bibliography, quoted material and small matches can be switched on or off. Two reports on the identical file can differ for that reason alone, before you even bring a free tool into it. Comparing a free checker's number to a department's number is comparing two measurements taken with different rulers.
Several writing assistants now bundle a plagiarism check alongside grammar correction. These are convenient and can catch obvious copy-paste from public websites, but they share the same structural limit as other free and consumer tools: they do not have access to the student-paper repositories that institutional checks search. A clean result from one tells you your text is not lifted from a public web page. It does not tell you it is clear of a senior's thesis.
There is also a side effect worth knowing about. The rewrite and paraphrase features in these same tools can change the character of your prose in ways an AI writing indicator responds to. Fixing a similarity problem with an automated rewrite can create an AI-indicator problem you did not have before.
Before uploading a thesis chapter to a free website, ask what happens to it afterwards. Many free tools say little or nothing about whether they store submitted text, how long for, or what it is used for. A free service still has costs, and the text people upload is sometimes part of how those costs are met.
For a short practice essay this may not matter much. For a thesis containing unpublished data, interview material from participants who were promised confidentiality, or an organisation's internal figures from an internship, it matters a great deal. Your ethical approval and your participants' consent did not extend to an unknown website.
There is a second risk in the other direction: if a tool does store and republish text in any form, your own work can later match against it. Avoid uploading complete theses to services you cannot vouch for.
The same caution applies to the people offering paid “Turnitin checks” through social media pages and cyber cafés. A check run on an account of unknown origin, with storage settings nobody can confirm, carries both risks at once: your file may be seen by people you never agreed to share it with, and it may be stored in a way that makes your final submission match against it.
Free checkers are not useless. They are just being asked to do a job they cannot do. Used for the right purpose, they are a reasonable first step.
For documents that matter, the useful thing is a check run against a database comparable to the one your department uses, with settings that do not store your file, and a person who reads the report with you. The number on its own is the least informative part. What helps is knowing which matches are citations doing their job, which are standard phrasing, and which genuinely need rewriting.
Related reading:
Not sure your free checker told you the whole story? Send the draft to us on WhatsApp. We will return a full plagiarism report and an AI writing report, run without storing your file, and explain what each match means before you submit.
Tell us what you need checked and when it is due. Your document is checked without being stored in any student-paper repository, so the check never counts against your real submission. The reports come back with what each match means, not just a percentage.