Set a paper, assign it to a class group, and mark the submissions — all on the school's own machines. Every submission carries an AI usage report, so a school can permit AI on one assessment and forbid it on the next and actually know the difference.
Setting, sitting, marking and analysing, in one place a teacher already has open.
Set a paper from the same workspace you build lesson material in, with deadlines attached. No separate exam authoring tool to learn.
Assign a paper to a specific group rather than to a list of individuals. Groups and their rosters are reused across every assessment you set.
Exams are stored and submitted locally, so a lab with filtered or saturated Wi-Fi is not a reason a paper cannot go ahead.
Open a submission, mark it in detail, leave comments, and request a resubmission where the work is not finished.
Generate class leaderboards and track improvement across a term, so progress is visible rather than reconstructed at report time.
See which AI models were used and how much, per student and per submission — the visibility a school needs before it permits AI in assessed work.
Students are notified when a paper is assigned and when feedback lands. Teachers see what has been read and what has not.
Export the whole assessment — scores, per-question detail, late flags — for your mark book or the school MIS.
Once generated text exists, every assessment that assumed unaided production needs a decision. The usual responses — ban it, ignore it, or buy a detector — all fail. Bans are unenforceable on personal devices, ignoring it invalidates the grade, and detectors produce false positives that fall hardest on students who write in a second language.
The workable answer is per-assignment rules plus visibility. Decide what help is permitted on each paper, state it, and be able to see what was actually used. That is why every submission here carries a usage report: which models were invoked and how much, per student.
The wider rollout sequence — assessment policy, data review, staff pilot, then students — is set out in our guide for schools.
The right tool depends on whether your constraint is network reliability or feature breadth. Grout is built for schools where the network is the constraint: exams are created, sat, submitted and marked on the local network, with class groups, AI usage reporting per submission and exportable results. If you also need to mark physical paper, the answer sheet evaluation module grades scanned handwritten scripts.
Yes. Papers and submissions are stored locally, so an exam does not depend on a live internet connection. This is the difference between a lesson that goes ahead and one that does not when the school network is saturated.
Every submission carries an AI usage report showing which models were invoked and how much. That visibility is what lets a school permit AI for some assessments and forbid it for others without relying on unreliable detectors.
Yes. Guardians linked to a student get progress reports covering exams, projects and homework, so parents see achievement without a teacher assembling a summary by hand.
No. Grout handles setting, sitting, marking and analysing the assessment, then exports the results. Your MIS stays the system of record for reporting and attendance.
That is a separate module. Answer sheet evaluation grades scanned or photographed handwritten scripts against a rubric you write, entirely offline. Most schools use both: digital exams for coursework, answer sheet evaluation for written papers.
Free download for Windows and macOS. Talk to us about volume licensing for a whole school.
Download Grout free