Mark the whole class in one go. Release nothing until you agree with it.
Attach each pupil's scripts, submit once, and come back to a marked class set — every mark grounded in the real exam board mark scheme, every comment quoting the pupil's own writing. You review, override what you disagree with, and release the lot together.
ReMarkAble turns a class set of work into reviewed feedback, a live markbook and evidence-led teaching actions — while the teacher keeps control of every judgement a pupil sees. For primary writing that means the same TAF language across the class, a live markbook, and nothing released until you agree.
Whole class in one submissionYou release, not the systemGaps ranked for teaching
The marking agent cannot answer without retrieving the actual exam board mark scheme — and we count whether it did, after the run, rather than taking its word for it. Across all marking from July to 23 September 2026: 898 pieces marked, 896 recorded as grounded and none recorded as ungrounded.
Quoted, then checked.
Every piece of evidence in the feedback is matched back against the pupil's own transcript. Of 7,288 quotes, 6,885 verified automatically.
Real coverage, not a demo.
86 live mark scheme stores, 264 papers and 1,637 seeded questions with their marks, sections and assessment objectives — across AQA, Edexcel, OCR, Eduqas and WJEC, plus KS1, KS2, 11+ and Years 1–6.
Quote verification is fidelity against the pupil transcript, not a marking accuracy percentage. We do not publish one.
How it works
1. Submit the class, not the script
Choose the paper. Attach each pupil's work — photographed handwriting or typed. Submit once. Every script marks in parallel and you can close the tab; anything that stalls is swept up automatically.
2. Moderate on one screen
The class set arrives in one moderation board, not thirty separate reports. Approve what you agree with, override the marks, grades and feedback you don't. The original AI judgement is kept alongside yours for audit. Nothing is visible to a pupil until you release it.
3. The markbook fills itself
Marks, grades, percentages and trends land per pupil per paper without anyone typing anything into a spreadsheet. Shared gaps across the class are ranked for you, with the pupils affected already grouped.
What it is for in primary
Not a replacement for teacher assessment. A way to mark a class set of writing, keep TAF language the same, and spend the time you save on teaching — with you still releasing every judgement.
Save marking time
Photograph the page. Confirm a proposed standard and next steps instead of writing the same comments after school.
Keep standards consistent
The same TAF language — WTS, EXS or GDS — with quoted evidence, so a year group is not marked five different ways.
See progress over time
Every piece is kept. You can see proposed standards and next steps accumulating across the term, not one snapshot in June.
Teach the remaining gaps
When the same TAF gap shows up across a set of pieces, you get a short lesson on that gap — not a generic worksheet.
Moderate the collection
STA judges a body of work, not a single good page. Propose one overall standard; a teacher agrees it.
File it and share it
Premium PDF for books, parents and internal moderation. The AI original is kept if you override.
From a photographed page to the next lesson
The writing below is STA 2018 exemplification — Kim in Year 2 and Leigh in Year 6 — so you can see the full loop on a public collection, not a pupil’s book.
Mark primary writing
Which year group?
123456
Photo of handwriting — ready to mark
1. Upload writing
Photograph the work, pick the year
Years 1–6. Handwritten pages are read as they are — no typing required. The year sets the National Curriculum expectations and the matching TAF, so the feedback is usable that afternoon.
KS2 Writing Assessment — Narrative
StandardEXS
KS2 Writing Assessment
Working at the Expected Standard
Narrative
Writes for a purpose with a clear opening and sequence.
Sentence demarcation is consistent enough for this standard.
Speech punctuation is the remaining next step.
2. Online report
Confirm the standard, keep the language consistent
A proposed WTS, EXS or GDS, strengths, next steps, and an evidence matrix against “pupil can” statements. Teachers confirm it. The same wording across a class makes moderation faster.
ReMarkAble AI
KS1 writing report
EXS
Working at the Expected Standard
Keep a copy for books, parents and the next time you look at this pupil’s writing. Quoted evidence stays with the piece.
3. A record of progress
Keep the trail, not just the last piece
Each marked page stays in the history. Premium PDF for books, parents and the next time you look at this pupil. You can see whether the standard and the next steps are moving.
On this Year 2 collection the remaining gap was tense consistency. On Year 6 it was punctuating direct speech. Each produced a 15–20 minute lesson — so the time you save on marking goes into teaching.
Writing is judged as a collection. Propose an overall standard across the pieces; a teacher agrees it. That is the conversation you would have in a moderation meeting — with the evidence already quoted.
The same loop, on writing you already know
Kim (Year 2) and Leigh (Year 6) are STA Expected Standard exemplification. We used them so you can see upload, report, lesson and moderation on a full collection — without putting a pupil’s work on a public page.
Year 2 · Kim · STA Expected Standard
KS1 Writing Assessment — Explanation
StandardEXS
KS1 Writing Assessment
Working at the Expected Standard
Explanation
Writes for a purpose with a clear opening and sequence.
Sentence demarcation is consistent enough for this standard.
Tense consistency within a sentence is the remaining next step.
Year 6 · Leigh · STA Expected Standard
KS2 Writing Assessment — Narrative
StandardEXS
KS2 Writing Assessment
Working at the Expected Standard
Narrative
Writes for a purpose with a clear opening and sequence.
Sentence demarcation is consistent enough for this standard.
Contains material developed by the Standards and Testing Agency for 2018 national curriculum assessments and licenced under Open Government Licence v3.0. View the licence. Source: KS1 Kim and KS2 Leigh on GOV.UK. Live marked packs: Year 2 Kim and Year 6 Leigh.
One shared language: WTS, EXS, GDS
Years 1–2 use the KS1 writing framework. Years 3–6 use the KS2 writing framework. That is the language teachers already use in moderation — so the report, the lesson and the collection judgement line up.
WTS
Working Towards the Expected Standard
EXS
Working at the Expected Standard
GDS
Working at Greater Depth
Primary maths, Years 1–6
Maths too, question by question
Photograph a pupil’s maths work and each question is marked, with the working read as it was written. Mark a whole class set at once and release the results when you agree.
Every question marked
Got it, Nearly or Not yet on each question, with whether the answer is right shown separately from the working.
One overall judgement
Working towards, Expected or Greater depth across the piece, in the language you already use.
The next teaching step
One next step chosen across the whole piece, so you know what to teach tomorrow.
Built to survive the questions you'll be asked about it
Every school leader buying AI assessment is really buying an answer to three questions: is it just guessing, who is accountable for the mark, and will the data stand up in a conversation with a governor or an inspector. Here is how ReMarkAble answers them.
Is it guessing?
The marking agent is required to search the actual exam board mark scheme before it can answer. The system then independently counts whether it did — it doesn't trust the model's self-report.
Every quoted line of evidence is matched back to the pupil's script. Unsupported quotes are detected and flagged, and a teacher reviews the result before anything is released.
Grade scales are enforced per qualification — GCSE 9–1/U, WJEC GCSE A*–G, A-Level A*–E/U, 11+ bands, primary Teacher Assessment Framework labels, and for primary maths Got it / Nearly / Not yet per question with an overall WTS, EXS or GDS. Curriculum for Wales writing is formative and gets no grade at all. A result that doesn't fit the tier is rejected and re-run, not rounded off.
Where the question context is ambiguous, it marks conservatively and flags it, rather than returning an off-task zero. Marking is best-fit, not defaulted down.
Who is accountable?
The teacher outranks the system. Full override of mark, grade and written feedback.
The AI's original judgement is preserved alongside the teacher's, so the moderation decision itself is a record.
Pupils see released rows only, and that is enforced in the database, not in the interface. Nothing is released by accident.
Will the data stand up?
One definition of every number. Attainment is calculated once, in one place, so your dashboard, your class report and your termly KPIs cannot disagree with each other about who was in the population.
Small cohorts are suppressed, not published. Comparison rows below three pupils and question-level gaps below four are withheld as noise — because a trend claimed off two children costs you more credibility than it ever buys.
Data we have no source for is shown as absent, not as zero.
Attainment breaks down by Pupil Premium, FSM Ever 6, SEN Support, EHCP, EAL, LAC and PLAC, service child, young carer, summer born, mid-year admission and prior attainment.
Safeguarding and data protection
Under-13 parental consent with verification, pupil data export, account deletion and administrator GDPR operations.
Role-based access enforced at the database layer — IT support and billing staff cannot reach pupil work. It isn't hidden from them; it isn't there for them.
Read-only "View As": a leader can see exactly what a teacher sees, with pupils pseudonymised on the server, and every access written to an append-only audit log. Acting as someone else is structurally impossible, not merely discouraged.
Reporting you get today: termly KPIs (assessments, pupils assessed, below expected, improving, active teachers), half-term progress series, attainment bands, class reports, the intervention lifecycle, and a report catalogue that states plainly what each report can and cannot produce. School-level reports export as CSV. Individual assessment reports and lesson plans export as PDF.
The part that actually gets your evenings back
One submission, one class.
Pick the paper, attach the scripts, submit. They mark in parallel and you can close the laptop.
One screen to moderate.
Not thirty PDFs. Override what you disagree with, release together.
A markbook that fills itself.
Marks, grades, percentages, trends — per pupil, per paper, no typing.
Bulk PDF export
for the whole class or a single batch. Parents' evening and books in one action.
Gap analysis you didn't have to do.
Shared weaknesses are ranked across the class, with the pupils affected, the severity and the supporting evidence. A gap has to reach half of an assessed cohort of at least five before we'll call it class-wide — because "your whole class needs this" off four scripts is how a tool loses a teacher for good.
Targeted lessons from real evidence.
Generate a teaching plan against the gaps your class actually has, then reassess.
An intervention that has a shape.
Identify, group, teach, reassess — with each group's stage visible instead of a flat list.
Independence recorded, and hidden from the marker.
Log a piece as independent, supported or heavily scaffolded for your moderation evidence. The marker never sees it, so it cannot bias the mark.
Registers that import.
CSV, columns in any order. Pupil accounts are optional — you can mark work for children who have no login at all.
Baselines, termly targets and roll-forward
between years.
What we won't claim
We don't publish an accuracy percentage.
No independent reviewer benchmark exists for our marking yet, so any number we gave you would be marketing. What we can evidence is that every mark is grounded in the real mark scheme and every quote is checked against the pupil's work — and we publish those figures above.
We don't claim to replace your judgement.
The system produces a first mark. You decide what a pupil sees. That is the design, not a limitation of it.
We don't yet report intervention impact.
We can show you the gap, the group and the teaching that followed. We cannot yet show you a clean before-and-after, so we don't pretend to.
We don't integrate directly with your MIS.
We import your register from CSV, including the usual MIS export formats. That is not the same thing, and we won't say it is.
“It's perfect for identifying gaps in writing early on. Our Year 1 pupils responded really positively.”
Miss Green — KS1 Teacher
Teachers remain the final authority on every judgement. Nothing reaches a pupil until you release it.
Questions leaders ask
See it on your own class before you decide anything
Fourteen days. Up to 30 pupils and 100 assessments. No card, no automatic billing, nothing to cancel. Mark a real set of your own scripts, moderate it, and look at what the markbook does with it.
If you want to see it working before you commit even that, ask for the demo school — a real class in use, with data already in it.