Writing the paper
Scope the chapters, set the marks scheme, get a draft in seconds. You still decide what goes out; you just stop starting from a blank page at nine on a Sunday.
Teachers
Set it Monday.
Set it Monday.
Reviewed by Tuesday.
Nobody became a teacher to mark forty scripts. The setting and the marking are the parts that eat a weekend; the reviewing is the part that needs you. This does the first two and hands you the third.
Set in minutes
Marked same day
Every score overridable
Logged with your name
Prints to A4
A paper, end to end
Monday to Tuesday,
Monday to Tuesday,
instead of Sunday.
The five moments a teacher actually touches. Everything between them is the platform’s problem.
Mon 4pm
Scope it
Pick the chapters, the marks scheme and the difficulty ratio. That is the entire input. If the chapters cannot fill the blueprint you get told which section fell short and by how much, rather than a quietly short paper.
- Chapters are a hard limit
- Marks scheme is yours
- You are told when it cannot be satisfied
Mon 4.08pm
Review the draft
Eight minutes with a generated paper. Change a question, swap one for your own, reorder a section. Nothing is published until you say so, and there is no auto-publish switch to find.
- Edit anything
- Type your own questions in
- No auto-publish, ever
Tue 9am
They sit it
Not your problem. Answers save on the device and sync when the line returns; a dropped connection costs a student the seconds it lasted. You watch a roster if you want to and ignore it if you do not.
- Offline-tolerant
- One screen for forty-two
- Grant time in one click, logged
Tue 12pm
Review the marking
Objective sections are already done. Written answers arrive marked against your rubric, each with the phrase in the student’s answer that earned the point. Twenty minutes for forty scripts, and you are checking rather than marking.
- Every score traces to a rubric line
- Half marks explained
- Override anything
Tue 1pm
Release
You decide when results go out. The moment they do, the chapter map exists — for you, for the students, and for their parents. Nobody has been waiting for a parents’ evening.
- Released on your word
- Chapter map, not a rank
- Practice issued automatically
Two of those five are yours. The other three used to be, and that is the whole proposition.
What it takes off a teacher
Six evenings back,
Six evenings back,
give or take.
Not features — the six specific jobs that stop being yours.
Marking the objective half
MCQ, fill in the blanks, true or false, match the following — done the moment the paper closes. That is usually half the marks and most of the tedium.
Marking the written half
Read against the rubric you set, with step marks for partial working and the phrase that earned each point quoted back. You review; you do not mark.
Building the analysis
The chapter breakdown builds itself from questions that were already tagged when the paper was drafted. No spreadsheet, no tally marks.
Chasing absentees
Who sat, who submitted, who never signed in — visible from the first minute rather than discovered at the end.
Telling parents
Results reach parents the moment you release them, with the chapter map. The conversation you have at the parents’ evening starts from something both of you have already read.
Where the weekend goes
Four ways to get
Four ways to get
forty scripts marked.
One paper, forty scripts, a class of thirty-eight. Scored on how much of it a teacher actually has to do. Click a row.
How the marking gets done
Teacher time in it
Score
Total
The unreviewed row is not the recommended one. Twenty-eight minutes of a teacher’s attention is worth more than any amount of model quality, and the product is built on that assumption rather than against it.
The honest bit
“I don’t trust
“I don’t trust
AI marking.”
Good. A teacher who trusted it immediately would be a teacher who had not read it carefully. Here is what we actually claim.
“I don’t trust AI marking.”You should not have to — you overrule it
Every score is a proposal, not a verdict. You can change any of them, and the change is logged with your name and a timestamp and appears on the result sheet. Nothing reaches a student until you release it, and there is no setting that skips you.
What we do claim is narrower: on objective sections it is exact, and on written answers it agrees with the reviewing teacher on most scripts and tells you which ones it was least confident about. Those are the ones to read first. It is a first pass that shows its working, not a replacement for your judgement.
“It will mark a good answer down for being unusual.”Sometimes — and it flags itself
This is the real failure mode, and it is more common than a wrong answer marked right. A student who reaches the correct conclusion by an unexpected route is exactly where confidence drops, and low-confidence scripts are surfaced first in your review queue for that reason. You will spend your twenty minutes on the four scripts that need you rather than the thirty-six that do not.
“My rubric is in my head.”Then the first paper takes longer, and it is worth it
Writing down what earns the second mark on a three-mark question is genuinely work. It is also the thing that makes two teachers mark the same script the same way, and most departments have never done it. The first paper is slower for that reason. Every paper afterwards is faster, and the marking stops drifting between whoever happened to pick up the pile.
“What about handwriting?”On-screen exams sidestep it; scans do not
If the paper is sat on screen there is no handwriting to read. If you are uploading scanned scripts, legibility is the limit and we will not pretend otherwise — poor handwriting produces low-confidence marks, which land in your review queue rather than in a student’s result. That is the honest boundary of what this does.
If those four answers do not satisfy you, the demo is thirty minutes and you can bring a script you have already marked and see where it lands.
Questions
What teachers ask
What teachers ask
in the first ten minutes.
The four that come up almost every time.
Yes, straight into any section. Type the question you want, set its marks, and it sits alongside the generated ones, counts toward the blueprint and lands in your department’s bank afterwards so the next paper can draw on it. Plenty of teachers use the generator only to fill the gaps around the questions they already knew they wanted.
Change it. The result updates, the student and their parent see the corrected mark, and the change is logged like any other override. There is no locked state that forces you to live with something you decided was wrong, and nothing is quietly rewritten behind a parent’s back — the log shows the correction happened.
Partly, and it is worth being clear about where. Anything with a defined right answer or a mark scheme works well. Extended essays, creative writing and open-ended argument do not, and we would rather tell you that than sell you a number you cannot defend. Most departments use it for everything except the subjects where marking is the teaching.
Not by us. The analytics compare divisions to each other and a cohort to its own prior year, because those are questions about teaching a chapter. There is no teacher leaderboard and no per-teacher score. What a head of department can see is which chapters a division lost marks on, which is a conversation about the syllabus rather than about a person.
Ready when you are
Bring a paper you already marked.
Book a free 30-minute demo. We’ll generate the same paper from your chapters and mark a script you have already marked yourself, so you can see exactly where it lands.