silvertext.
LoginSign up

Privacy

Do plagiarism checkers store your work?

Most of them keep it. Far fewer share it. Of the ten services below, one publishes that your text joins a corpus other people are later matched against, and one more offers it as a setting. Seven keep a copy of your text so you can reopen your report. Another three publish nothing at all on the sharing question, and a silence is not a no.

Every line below is the vendor’s own published wording, quoted and linked. This is not a ranking, nothing here is scored, and no accuracy claim appears anywhere on the page.

Sources read

Of 10 services

1
share your text for matching
1
offer it as a setting
7
keep a copy of your text
3
publish nothing either way

Three questions get asked as one

Almost every confusing answer in this market comes from a vendor answering one of these and letting you hear all three.

  1. 01

    Does it join a corpus other people are matched against?

    This is the one most people mean, and the only one that can make your own words come back flagged when you or somebody else runs a check later. It is also the rarest to publish plainly.

  2. 02

    Is a copy kept at all?

    Almost every service keeps your text, usually so you can reopen your report. Kept is not the same as shared, and a service can be honest about one while saying nothing about the other.

  3. 03

    Is it used to train a model?

    A newer question than the other two, and the one most policies were not written to answer. Some say yes with an opt-out, some say no, most say nothing.

At a glance

What 10 services publish

Alphabetical, Silvertext included in its alphabetical place. “Not stated” means the vendor is silent, which is different from a no.

  • Copyleaks
    Shared for matching
    Opt-in, off by default on the API
    Text kept
    Yes, the terms reserve the right
    Training
    Not stated
  • Grammarly
    Shared for matching
    No, per its product page
    Text kept
    Not stated
    Training
    Account-wide, opt-out exists
  • Originality.ai
    Shared for matching
    No published position
    Text kept
    Yes, in your dashboard
    Training
    Yes, unless you opt out
  • PlagiarismDetector.net
    Shared for matching
    Says nothing is stored
    Text kept
    Says no
    Training
    Not stated
  • Quetext
    Shared for matching
    No, not the searchable database
    Text kept
    Yes, as a report in your account
    Training
    Allowed by its policy
  • QuillBot
    Shared for matching
    No published position
    Text kept
    Yes, so you can reopen reports
    Training
    Not stated
  • Scribbr
    Shared for matching
    No public or shared database
    Text kept
    Yes, in a private database
    Training
    Not stated
  • Silvertext
    Shared for matching
    No
    Text kept
    Yes, until you delete it
    Training
    No
  • Smodin
    Shared for matching
    No published position
    Text kept
    No published position
    Training
    No published position
  • Turnitin
    Shared for matching
    Yes, on the general repository setting
    Text kept
    Yes
    Training
    Not stated
  • Yes, it does
  • Only on a setting
  • No, it does not
  • Not published

Service by service, in their own words

Copyleaks

Shared for matching
Opt-in, off by default on the API
Text kept
Yes, the terms reserve the right
Training
Not stated

“Add the submitted document to the Copyleaks Shared Data Hub.”

Copyleaks API reference, no date shown on the page

There are two databases here, not one. The Shared Data Hub, which other customers can be compared against, is documented as off by default on the API. A separate per-account index, which only your own later scans are compared against, is documented as on by default. Copyleaks does not publish the default for its browser app or its LMS integrations, so do not assume the API default carries over. Its terms of use, last updated 24 August 2026, separately reserve the right to keep uploaded documents in a database.

Grammarly

Shared for matching
No, per its product page
Text kept
Not stated
Training
Account-wide, opt-out exists

“never publicly searchable or stored in any database accessible by others”

Grammarly plagiarism checker page, no date shown on the page

Read that sentence closely. “No database accessible by others” is a promise about who can see your text, not a promise that it is deleted. Grammarly does not publish a retention position for checked text anywhere we could find. Its privacy policy now redirects off grammarly.com to a page branded Superhuman Platform Inc. (formerly Grammarly), effective 6 July 2026, which says information collected is used to train its AI models and points to a training control in account settings. Neither that policy nor the support article for the control mentions the plagiarism checker, so the training answer here is the account-wide one and not a statement about checked text.

Originality.ai

Shared for matching
No published position
Text kept
Yes, in your dashboard
Training
Yes, unless you opt out

“use your scan history and scan results to help train and improve our models”

Originality.ai privacy policy, dated 17 June 2026

The training position is the clearest of any service here, and it is opt-out rather than opt-in. On whether scans join a corpus other customers are checked against, Originality.ai publishes nothing either way, and we are not going to read that silence as a no.

PlagiarismDetector.net

Shared for matching
Says nothing is stored
Text kept
Says no
Training
Not stated

“We do not save or store the user content uploaded for checking plagiarism.”

PlagiarismDetector.net privacy policy, no date shown on the page

The strongest claim on this page, and the only one with no date attached to it. The policy carries no effective or last-updated date, so there is no version of it you could point back to later. We read it on 18 September 2026.

Quetext

Shared for matching
No, not the searchable database
Text kept
Yes, as a report in your account
Training
Allowed by its policy

“Your data does not become a part of our searchable database.”

Quetext help center, dated 24 October 2025

Two separate things, and the difference matters. Your text does not join the corpus other people are matched against, and it is kept in your account as a report. Quetext’s privacy policy, last updated October 2025, separately says user content may be used in its AI training, tuning and evaluation work.

QuillBot

Shared for matching
No published position
Text kept
Yes, so you can reopen reports
Training
Not stated

“only uses your uploaded content to run a plagiarism report”

QuillBot help center, dated 10 June 2026

The article says the content is used to run the report and to let you open that report again later, which means it is kept. It makes no claim either way about a corpus other people are checked against.

Scribbr

Shared for matching
No public or shared database
Text kept
Yes, in a private database
Training
Not stated

“you give Scribbr permission to save the Document in a protected and private Database”

Scribbr terms of use, dated 21 August 2026

Both halves of this row are Scribbr’s own words and they are not in conflict once you separate the two questions. Its help center, updated 23 March 2026, says the document is not stored in any public or shared database. Its terms say the document is saved in a private one. Public and private are the point.

Silvertext

Shared for matching
No
Text kept
Yes, until you delete it
Training
No

“It is not added to any archive that other people are checked against”

Silvertext integrity policy, dated August 2026

Hold this row to the same standard as the others. The file you upload is deleted once its text has been read; the text itself stays in your report so you can come back to it, and it goes when you delete the report or the account. We do not train on it, and we configure our providers not to train on it wherever they offer that control, which is a narrower promise than “never, anywhere” and is written that way on purpose.

Smodin

Shared for matching
No published position
Text kept
No published position
Training
No published position

No published statement to quote.

Smodin legal page, dated 30 March 2026

Its legal page covers account and personal data. We could not find a sentence on any Smodin page that addresses what happens to the text you put into the plagiarism checker, so there is nothing here to quote and nothing to hold them to.

Turnitin

Shared for matching
Yes, on the general repository setting
Text kept
Yes
Training
Not stated

“submissions are stored in a general Turnitin repository”

Turnitin assignment settings guide, no date shown on the page

This is the one that answers the question most students are actually asking, and the setting is not theirs. The instructor chooses between the general repository, an institution-only repository, and not storing the submission at all. Turnitin’s own privacy policy, dated 4 February 2026, does not describe the repository; the product guides do.

How to check any checker in four minutes

This page will go stale. The method will not.

  1. 01

    Read the terms, not the FAQ

    An FAQ is written by marketing and a terms page is written by lawyers, and when they disagree it is the terms you agreed to. Two of the services here read differently depending on which page you open.

  2. 02

    Search the page for three words

    Database, retain, train. A policy can answer one of those clearly and be completely silent on the other two, which is the most common shape by far.

  3. 03

    Check whether the page is dated

    An undated policy cannot be held to a version. If a claim matters to you, note the date you read it, because that is all you will have later.

  4. 04

    Ask who owns the setting

    With a checker your school runs, the storage choice is usually the institution’s or the instructor’s, not yours. Knowing the policy does not mean you control it.

FAQ

What people ask next

Still unsure how this applies to you? Ask us. A person answers.

Will checking my own paper make it come back as plagiarized later?

Only if the service adds your text to a corpus that later checks are compared against. Of the ten here, one publishes that it does on its general setting, and one offers it as an option. The rest either publish that they do not, or publish nothing.

Can I stop my paper going into a repository at school?

Usually not directly. With Turnitin the repository choice sits in the assignment settings, which belong to the instructor or the institution, and the options are a general repository, an institution-only one, or no storage. Asking your instructor which one the assignment uses is the only reliable route.

Is “we do not put it in a database” the same as “we delete it”?

No, and the gap between those two sentences is where most of the confusion on this topic lives. Several services here state clearly that your text is not in any shared database, and none of those same statements say the text is deleted.

Does it matter if a checker trains on my writing?

It depends what you are protecting. Training on your text does not make it retrievable by another user the way a matching corpus does, but it is still a use of your work that you may not have intended, and it is the use most often buried in an opt-out rather than offered as a choice.

What does Silvertext do with the text I check?

Your text is not added to any archive other people are checked against, and we do not use it to train models. An uploaded file is deleted once its text has been read, the text stays in your report so you can come back to it, and it goes when you delete the report or your account. The integrity policy is the binding version of all of that.

How current is this page?

Every source here was read on the vendor's own domain on 18 September 2026, and each row shows the date that vendor's page carried, or says plainly when it carried none. Policies move, so check the linked source before relying on any row.

Our row is in the same table

Silvertext does not add your text to any archive other people are checked against, and does not train on it. That is a claim like every other claim on this page, so hold it to the same standard: the integrity policy and the privacy policy are the versions that bind us, and they are worth reading before you take our word for it here.