Partly, and probably not in the way you are worried about. Turnitin's crawler indexes public web pages, so anything from Chegg or Course Hero sitting on an open page can turn up as a match. What it cannot reach is content behind a login or a paywall, which is most of what students actually use.
That makes the honest answer to the headline question "sometimes, at the edges." It also makes it the wrong question, because matching software is not the mechanism that catches people using these sites. A separate process is, and almost nobody knows how it works.
What the crawler can and cannot see
A plagiarism checker can only match against documents in its index, and its index is built by crawling pages it can reach. Public Course Hero previews, forum reposts, and anything either site allows search engines to see are all fair game.
Content that requires a subscription is a different matter. A crawler that cannot log in cannot index what is behind the login, so a solution you opened with a paid account is not sitting in the comparison set waiting to match your submission.
There is one exception, and it catches the people who assume a paywall protects them. Material moves. A solution posted on Chegg gets copied to a public forum, or a blog, or somebody's study-guide site, and the moment that copy is indexed the text is findable no matter where it started. You have no control over that and no way to check it beforehand.
The route that actually finds people
Institutions do not rely on similarity matching for this. They ask the platform directly.
Chegg's honor code page describes the process: an institution submits an official request on letterhead signed by a dean or academic integrity officer. Chegg states this requirement cannot be waived, on student-privacy grounds. So an individual instructor who suspects something cannot simply email and get an answer, which is why these investigations run at the department level or above.
What Chegg says it provides in response is narrower than the campus rumour suggests: for each URL submitted, the timestamps for when the question was posted and when the solution was posted. Personal user information is withheld. Violating accounts may be terminated and material removed.
Read that carefully, because it cuts both ways. Chegg is not handing your name to your dean on request. But a timestamp showing a question posted eleven minutes into a two-hour exam is not an ambiguous document, and the institution already knows who was sitting that exam.
Where the identification actually comes from
The timestamp establishes that it happened. Connecting it to a person is the institution's job, and universities have their own routes for that.
The University of California, Irvine's guidance on contract cheating is unusually explicit. Faculty can obtain student access logs from third-party services and ask campus IT security to correlate that data with university login records, following the institution's process for non-consensual access to electronic communications. The office also issues signed letters of support so faculty can submit investigation requests to Chegg and copyright reports to Course Hero.
That is the actual pipeline: platform confirms the event and its timing, the institution links it to an account, and the case is built from the correlation rather than from a similarity percentage.
It also explains why these cases arrive in batches instead of one at a time. An instructor who finds their exam posted has found one URL and one timestamp, but the request covers everyone who touched that question, so a single discovery can open a dozen files at once. Students describe this as the department suddenly cracking down. What happened is that one query returned a list.
Honor Shield, and why the exam question disappeared
There is also a preventive tool. Honor Shield lets verified educators submit exam questions in advance so that solutions to those questions are blocked on the platform during the testing window. It is free, and instructors are told to upload at least 48 hours ahead.
Its limits are worth knowing, because they explain inconsistent experiences. It cannot block answers that already exist in the textbook solutions database, and questions posted outside the specified window are not covered. So a student may find a question unanswerable during one exam and freely available during another in the same course.
Course Hero runs on a different lever
The Chegg route is an integrity investigation. For Course Hero, the mechanism institutions reach for first is copyright.
Course materials belong to somebody. An exam paper, a problem set, a lecture slide deck: those are typically the work of the instructor or the institution, and uploading them to a study-sharing site is a distribution of someone else's material. UCI's office states plainly that it supports faculty in submitting copyright infringement reports to Course Hero, which is a faster and better-established process than an integrity request because every content platform has a legal obligation to handle takedowns.
This is why students are sometimes surprised by which upload causes trouble. Posting your own completed essay is your material to post, whatever your institution thinks of it. Posting the exam paper it was based on is somebody else's, and that one has a legal route attached that does not depend on any academic policy at all.
The other detection layer: how you write
Separate from all of this, Turnitin markets an authorship tool that compares a submission against a student's previous work, looking for inconsistencies in writing style rather than matching text against sources.
Treat that as context rather than something to fear precisely. Style analysis produces a signal, not proof, and its output is a prompt for a human conversation. But it means the assumption that work is safe if it matches nothing in an index is out of date. Nothing needs to match for a submission to look unlike everything else you have written.
What this means in practice
If you are wondering whether you are exposed, the useful questions are not about the similarity checker:
- Did you post an exam question during an exam? That is the one with a clean timestamp attached and the highest chance of institutional action.
- Did you submit a solution as your own work? Then whether it matches anything is beside the point. The problem is the submission, not the detection.
- Did you read a worked example and then write your own answer? That is studying. It is what the sites claim to be for, and it produces no timestamp, no match, and no case.
The distinction that matters is between using these sites to understand a problem and using them to produce a submission. Detection technology is a poor guide to which side of that line you are on, and the line is the thing your institution's policy actually describes.
The short version
Turnitin will occasionally match Chegg or Course Hero content, when that content sits somewhere public or has been copied there by someone else. Relying on the paywall to hide it is a bad bet for that reason.
But if you are trying to work out your risk from a similarity report, you are reading the wrong instrument. The cases that go somewhere start with a dean's letter, a timestamp, and an access log, and none of those appear in the coloured highlights on your similarity score. If your reason for asking is simply that you want to know what a marker's software will see in your own writing before you submit it, that is what a plagiarism checker is genuinely good for.