Is my paper stored in Turnitin forever?
Usually your paper is kept, and you did not choose that. Storage is one setting on the assignment, picked by your instructor from the options your institution permits, and there are three: kept in the global repository, kept where only your own institution can search it, or not kept at all. Turnitin's user agreement puts that decision with the institution in as many words. Removal is possible and properly documented, but the request travels through your instructor and your institution's Turnitin administrator, never through you and never through Turnitin's support team. As for forever, no Turnitin document names a period after which a stored paper is deleted, and the licence you accepted has no end date.
Most people who type this question are asking a narrower one underneath it: will the essay I handed in last term come back as a match against the one I am writing now. That has its own rules and reusing your own work covers them. This page is about the storage itself. What the three settings do, who picks between them, what somebody at another university can see, and what it takes to get a paper back out.
What stored actually means here
The assignment your instructor built has a setting called Submit papers to. Turnitin's own instructor guide lists three values and describes each in a single line.
| Option, classic assignment | Same option, newer assignment type | What Turnitin's guide says it does |
|---|---|---|
| Standard paper repository | General Turnitin repository | "Submissions are stored in the global Turnitin repository" |
| Institution paper repository | Private institutional repository | "Submissions are stored in a repository which is only accessible by your institution" |
| Do not store the submitted papers | Do not store the submitted papers | "A Similarity Report will be generated for submissions, but submissions will not be stored in the standard repository or the institution's repository for future comparison" |
Both naming schemes are live, because Turnitin runs an older assignment type and a newer one side by side, and the classic settings guide is worded differently from the newer one. Two students comparing notes can describe different menus and both be right.
Now the part most pages on this question get wrong. Storing and comparing are separate settings. Directly below Submit papers to sits a second block called Compare against, and Turnitin describes them independently: the comparison option "compares current submissions to previous submissions in classes and assignments in Turnitin across institutions", while the storage option decides whether yours joins them. An assignment can check your work against every stored paper without adding yours to the pile, or add yours while comparing against almost nothing.
"Is my paper checked against other students' work" and "is my paper kept so it can be checked against theirs" are two different settings on the same screen. Turnitin's own documentation separates them. Almost every explainer on this topic collapses them into one.
The setting is your instructor's, and sometimes not even theirs
There are three layers of control here and you are not on any of them.
At the top, an administrator at your institution decides which repository options instructors are allowed to see at all. Turnitin's account settings guide lists three: one lets instructors choose between the standard repository and no repository, a second adds the institutional repository to that choice, and a third is Submit all papers to the standard repository, described as "All student papers submitted to the account will be stored in the standard paper repository". At an institution configured that way, your instructor has no choice either.
In the middle, your instructor picks a value per assignment. Turnitin's advice on when to pick the third option is unglamorous: it suggests selecting no repository when "creating a new assignment (such as for rough drafts, practice submissions, or testing)". That is a teaching decision about draft work, not a favour a student can request. Asking for it in order to stay out of the index is the wrong question put to the wrong person, and it reads exactly the way it sounds.
At the bottom is you, and the honest finding is that there appears to be nowhere in the student view where the setting is shown. Turnitin's guide to submitting to a Standard assignment walks through the whole student flow without using the word repository once, and its help article for students who want a paper deleted assumes you will have to ask. If your institution publishes its Turnitin configuration, read that. Otherwise the only documented way to find out is to ask the person who set the assignment.
The user agreement you accepted the first time you submitted states the arrangement plainly: "Submitted papers are held at the discretion of Customers in accordance with the terms of the Registration Agreement between Turnitin and Customer." Customer means your institution. The same section adds that you may "at any time request that Turnitin and its service providers delete each paper that you individually identify to us through the system administrator at your institution". That is the removal route, and it runs through your university.
What a stored paper does to your next one
A stored paper is not a verdict sitting on a file. It is a document that later submissions get compared against, and the comparison rules are more forgiving than the folklore suggests.
Inside one class, Turnitin excludes your own earlier work automatically. Its help page on self-matching says that if you submit the same paper to more than one assignment in the same class, "Turnitin automatically prevents these papers from matching against each other", and explains why: the default self-match exclusion "is designed to allow students to submit early drafts of their work without their final submissions matching against their own previous drafts". Those hidden matches are not gone. They sit in the Excluded sources section of the report, "where they can easily be re-added to the report if necessary" by the instructor.
Across different classes, the automatic exclusion stops. Turnitin is explicit: an automatic exclusion "does not occur" when the same paper goes to assignments in different classes, "provided at least one of those assignments was configured to save papers to the standard Turnitin repository". Instructors can still exclude the match by hand once they see what it is.
So a resubmission of your own work in a new module produces a match, and the match names you. That is a conversation about whether reusing the work was permitted, which your course rules answer and Turnitin does not. The number attached to it means less than people assume, and what counts as a good similarity score goes through why a percentage on its own decides nothing.
What a marker at another institution actually sees
This is the part almost nobody documents, and it is the part students most want to know. Being in the global repository does not mean strangers can read your work.
When a submission at another university matches yours, that instructor sees the match, its size, and a label reading Submitted to followed by the name of your institution. They do not see your text. To see it they have to ask, through a mechanism Turnitin calls a paper view request, and Turnitin explains why it works that way: "All intellectual property rights remain with the student when they submit their paper, making Turnitin unable to offer immediate access to papers or handle the paper exchange directly."
What happens next is three steps in Turnitin's own words. "We send an email on your behalf to the host institution. An instructor or privacy officer at the host institution receives the email request. If approved, the original paper will appear in the reply email." The email arriving at your institution carries a recap of the assignment, the requesting instructor's name and the similarity percentage, plus "a full plain-text version of the original paper, as it was submitted to your class". Whoever receives it decides, and Turnitin's instruction to them is unambiguous: "While Turnitin encourages instructors to collaborate and work together with paper view requests, you are under no obligation to respond." If they do not, "Turnitin cannot pursue a paper view request any further."
The person holding your work is therefore at your own institution, not at the one asking, and a request that gets ignored simply ends, because there is no escalation path.
Inside a single institution the rules are looser and an administrator controls them. Turnitin's account settings include a Submission Release feature which "allows other instructors associated with your institution to view papers submitted to your classes without having to go through the paper view request process". Whether your university has switched that on is a question for its own policy documents.
Taking a paper back out
Deletion is real, documented, and slower than people expect. It is also entirely internal to your institution.
Turnitin's instructor guide states the boundary first: "Submission and composition deletions cannot be requested from Turnitin Support and can only be managed by your institution's administrator." Its student-facing version says it to you directly: "For contractual and data privacy reasons, Turnitin cannot process deletion requests directly from students." Emailing Turnitin achieves nothing, which matters because it is the first thing most people try.
The documented path has four moving parts.
- Find your Submission ID. Open the paper in the document viewer, click the information icon in the right-hand sidebar, and copy the Submission ID. Nothing can be found without it.
- Give your instructor three things. Turnitin tells students to supply the Submission ID, the class name and the assignment name. Send it in that shape so nobody has to go looking.
- Your instructor raises the request. From the assignment inbox they choose Request permanent deletion, pick a reason, and type DELETE to confirm. A trashcan and clock icon then marks the paper as pending.
- An administrator approves or rejects it. They can act on requests one at a time or approve up to ten at once, and can delete a submission directly without any request at all.
If it is approved, Turnitin's guide says the submission is removed "from everywhere in Turnitin, including from the instructor's and student's assignment and Turnitin's repositories", along with "all generated reports and any associated grading and feedback", and it "will not match against future submissions". Reports that already show it need refreshing. This is irreversible: "Deleting a submission is permanent and cannot be recovered."
Two things the guides bury. Deleting a paper also deletes the grading and feedback attached to it, which is a real cost if you wanted either. And rejections leave a trail: administrators have an activity audit covering both deleted submissions and rejected deletion requests, recording who asked, for what, and when.
How long, in Turnitin's own words
The honest answer to the headline is less satisfying than either of the confident versions circulating online.
Turnitin's services privacy policy has a section headed Data Retention and it runs to two sentences. Personal information may be retained "for the period necessary to enable the continued use of the Services", and "How long we retain specific Personal Information varies depending on its type and use, after which it will be deleted." No period is named. Not for submissions, not for anything else.
The user agreement is specific about the shape of the arrangement, though not about time. You keep ownership of the paper. Turnitin gets "a non-exclusive, royalty-free, perpetual, worldwide, irrevocable license to use such papers" for two stated purposes, providing the service and improving it, and adds in bold that "Any cessation of the use of the Site or Services shall not result in the termination or expiration of this license." Graduating does not end it. Closing the account does not end it.
So nothing on record says forever, and nothing on record sets a date. What is on record is a licence with no expiry and a deletion process with no clock, which means a stored paper leaves the repository when somebody removes it and not before. Nobody publishes how often each of the three settings is used, either. Turnitin does not, and we found no institutional survey that measures it.
If you are in the UK or the EU there is one genuine right here. Turnitin's GDPR material confirms that "Under the GDPR, submission content itself is 'Personal Data'" and that its database of submissions "is located in the United States". On rights of access, deletion and rectification, the privacy policy gives one instruction: "To exercise your rights, please contact your education institution. We will work with them to facilitate your request in circumstances in which they require our assistance." Turnitin describes itself as a data processor under the EU and UK regulations, with your institution as the controller, and its GDPR FAQ adds that "Generally, the customer, as data controller, must approve (or deny) such requests". The right is real, it is exercised against your university, and your university decides.
What this changes about how you work
Almost nothing, which is the useful conclusion. A stored paper is a record of what you handed in on a particular day. It becomes visible again only if the same text turns up in another submission, and within a module it usually will not. There is no retroactive judgement and no score attached to you.
What does change is that similarity is only one of several things being run over your work, each with different rules and different reliability. The AI writing indicator is a separate subsystem from the repository and an administrator can switch it off entirely, which is why what a Turnitin AI score means and the universities that have turned AI detection off answer a different question from this one. For what these tools can actually establish, what Turnitin can and cannot detect is the honest version, and the habits that keep you clear of all of it are in how to avoid plagiarism.
If you have a real data protection concern, and some people do, put it in those terms. Work containing personal information about you or somebody else, or a paper you intend to publish and do not want indexed, is a legitimate reason to write to your institution's data protection officer and ask what its Turnitin configuration is. That is an ordinary request, answered by the people who chose the setting. It is not a route to keeping work out of the index because you would rather it were not compared, and nobody involved will read it as one.
None of this is about writing. It is about knowing where your work has gone, which is the same instinct that makes people check what a source actually says before citing it. That is the part CiteOwl is built for: it searches OpenAlex, CrossRef, Unpaywall and the web, reads the papers it retrieves, writes each claim from a passage it keeps beside the sentence, and hands every edit back as a diff you accept or reject. What ends up in your reference list is what a real record said, and you saw it arrive.