AI detectors flag honest students and miss dishonest ones: they guess from the finished text and get it wrong in both directions. Folio Classroom takes the opposite approach. Instead of a probability score, you get positive evidence of how each document was actually written. One link in your syllabus, and the rest follows.
It's free for instructors.
Create an assignment
Open Assignments from the menu under your avatar and choose New assignment. You'll set:
- Title: what students see (e.g. "Essay 2: Argument analysis"). You can also start from a template.
- Instructions (optional): any brief the students should read before they start.
- Due date and time (optional): you set it in your time zone, and each student sees it on their own clock. Submissions after it are flagged late on your roster. You can give an individual student an extension.
- Close submissions after (optional): after this moment Folio refuses new submissions and resubmissions, except from a student whose extension runs later.
- Students submit a: Document (written in Folio, with a full Integrity Certificate) or Video (an uploaded recording, for an oral exam or a presentation).
- AI-use policy: state the rule for this assignment, or leave it with no stated policy.
- Reviewer rubric (optional): one of Folio's built-in rubrics; students are told to self-review against the same dimensions you grade on.
- Grading criteria (optional): your own dimensions, in your words, optionally with performance levels so grading is one click per criterion. Students see them under "How you'll be graded".
- Grading: Hide student names while I grade (blind grading) and Hold grades until I release them.
- Course (optional): adding it to a course lets enrolled students find it and rolls its grades into the course gradebook.
Create & get the link, and that's the entire setup. Put usefolio.co/a/YOURCODE in your syllabus, your LMS, or an email.
How students submit
A student opens your link, signs in (a free account takes under a minute), and joins. Joining shares their name and email with you, and the page tells them so before they commit.
They then pick a document they've written in Folio and submit it. Submitting publishes the document's Integrity Certificate; that is the submission. Each certificate records:
- word count and the number of saved revisions,
- how many sources it cites, with each source's standing (a resolvable DOI, and whether it has been retracted) shown on the certificate's verify page,
- every AI-assist action, disclosed,
- measurements of the writing process from the edit history: how many sessions and days it was written over, and how much of the final text the largest single save added, and
- a tamper-evident fingerprint of that history.
A submission is frozen. Folio keeps a copy of exactly what was submitted, and that's what you read, grade and download, even if the student edits, trashes or revokes the document afterward (the roster says when that happens). Students can resubmit until the assignment closes. A resubmission refreshes the same certificate and starts a new, ungraded attempt; the grade and feedback on each earlier attempt stay in the roster's history.
Students get the strongest record when they draft in Folio, where the process is captured as it happens. They can import existing work, but the certificate attests to what Folio actually saw.
A Video submission gets a different kind of record, since there's no writing process to attest to: Folio signs a video upload receipt instead of an Integrity Certificate. It names who submitted it and when, and records the stored file's size, format, and a storage fingerprint, so anyone with the link can confirm the file on record hasn't changed since upload. It's honest about what that does and doesn't show — it doesn't prove who was on camera or that the recording is unedited.
Read, grade, and export from one roster
Open any assignment to see its roster. At the top, a quick summary: how many submitted, how many on time, how many graded, and how many submissions had 90% or more of their final text arrive in a single save. That is a fact about how the text was saved, worth a closer read beside the certificate, not an accusation. Folio does not give a writing-process score or a label about how a student wrote.
For each submission you can:
- Read it in place: a clean, read-only view of the version the student submitted, right on the roster. To leave comments in the text itself, use Comment in document, which opens the student's document read only. Nothing you do there can change their text.
- Grade it: a grade in your own scheme (
A-,17/20,Pass) plus written feedback. If you aren't holding grades, saving notifies the student, who sees the grade and feedback on their assignment page. If you are, students see "graded, pending release" until you publish grades to the class, now or on a date you schedule. - Download it as a
.docx, references included, or download its authorship and process Report as a PDF. - Verify the certificate (or, for a video submission, its upload receipt): the same public link anyone can check.
Grading usually happens after you close an assignment, and that's fine: students keep seeing their submission and grade after it closes.
Take everything to your gradebook
Exports live at the top of the roster:
- CSV: the whole roster (names, word counts, saved revisions, whether most of the text arrived in one save, grades, feedback, late flags, certificate links) for import into any gradebook.
- All .docx (.zip): every submission as a Word file named Student Name, Document Title.docx, ready for offline grading or upload to Moodle, Canvas, or wherever you already work.
- Class report: a shareable PDF summary of participation, grades and the writing-process signal. It prints Arabic and Hebrew names and text correctly, right to left where they should be.
Why this beats an AI detector
A detector asks "does this look AI-written?" and punishes the students whose honest prose happens to look tidy. Folio asks "how did this document actually come together?" and shows you the answer: the revisions, the sessions, the sources, the disclosed AI use. The one flag is a measurement: 90% or more of the final text arriving in a single save. Even then it's a prompt to read more closely, not a verdict. Everything else is the evidence a careful writer wants on their side.
Already collected work elsewhere? You can verify any batch of certificate codes at usefolio.co/check.
Related: Peer review in a course