I was stuck on a 5 for ages. The AO2 feedback was so specific I knew exactly what to fix. Got a 7 in my mock the next week.
M
Maya O.
Year 11 · AQA
LIVEUK GCSE · AQA · Edexcel · OCR · WJEC
Turn Grade 5 essays
into Grade 7 answers.
Write your answer. Our deterministic AI examiner returns a real GCSE grade, AO-mapped feedback, and exact next-grade steps — calibrated to your exam board.
No card neededGrades 1–9 in secondsAO1 · AO2 · AO3 mapped
92%
Within ±1 grade
4.2s
Avg. mark time
50K+
Answers marked
marked.app — AI GCSE Examiner
AQA English Literature · Jekyll & Hyde · 20 marksHow does Stevenson present Hyde as a frightening character?
AO2 DetectedAI examiner ready
GCSE Estimate
Awaiting marking…
AO1
82%
AO2
88%
AO3
71%
Strengths
- 1Quotations are embedded and analysed for effect, not feature-spotted
- 2Context is linked to reader response
How to reach grade 8
Offer two interpretations of a key quotation and link Hyde’s regression to Darwinian anxiety — that is what separates Grade 8 analysis.
Calibrated against UK exam boards
AQALargest UK board
EdexcelPearson
OCROxford · Cambridge · RSA
WJECEduqas
iGCSEInternational
CIECambridge
AQALargest UK board
EdexcelPearson
OCROxford · Cambridge · RSA
WJECEduqas
iGCSEInternational
CIECambridge
How it works
Three simple steps to a better grade
No complicated setup. Just practice — the way your exam actually works.
01
Write your answer
Pick a real GCSE-style task or paste your own. Write it like you would in the exam — no templates, no fluff.
02
Get instant AI marking
Our deterministic engine grades against AO criteria — returning a 1–9 grade, AO breakdown and feedback in seconds.
03
Improve to the next grade
See specific strengths, exact fixes, and what to add next. Rewrite, re-mark, watch your grade climb.
0%
Within ±1 grade of human examiner
Calibrated against agreed reference set
0s
Average AI marking time
From submit to AO-mapped feedback
0+
Practice answers marked
Across Years 7–11 in the UK
0 grade
Average improvement after 4 weeks
Tracked via dashboard progression
Under the hood
The deterministic marking engine
Not a prompt wrapped around a chatbot. AI gathers the evidence; a pure-code engine anchored to real exam-board mark schemes decides the grade — and locks it before a single word of feedback is written.
Question routing
Pure code · runs first
Not every question is an essay. Each one is routed to the marking method its real mark scheme uses — exact-match for multiple choice, point-marking for short comprehension, level-of-response for extended writing. Anything we can't mark honestly is reported as unmarked, never given a fabricated zero.
objectivepoint-markedlevel-markedunmarkable
Answer screening
Pure code · no AI, no wait
Every answer is screened before it reaches the engine — off-topic drift, keyboard mash, symbol spam, and text pasted straight back out of the source extract. Students see the warning while they can still fix it, rather than a zero afterwards.
off-topicgibberishcopy-overlaplength
Only what survives both checks is worth spending a model on
01
Feature Extraction
AI call · analytical only
The first AI call is never shown a grade scale. It reads the answer against the real question — including its source texts, insert extracts and described images — and returns measurable signals as structured JSON.
ideaClaritystructurevocabularyRangetechniquescoherencegrammarErrors
02
Examiner Judge
AI call · descriptors only
An independent second pass reads the answer against the exam board's own published level descriptors and reports which level it meets, how confident it is, and what is holding it back. It returns a level — never a grade.
awardedLevelconfidencewhatPreventsHigher
Two independent passes — neither one returns a grade
03
Deterministic Grade Engine
Pure code · zero AI calls
Everything above hands back evidence. This layer decides — in code, with no model in the loop. The same evidence produces the same grade today, next term and next year.
AO-weighted scoring
Per-board, per-task assessment-objective weightings live in the database as data — never as branching logic in code.
Descriptor anchoring
Where a board publishes a real level grid, the awarded level sets the grade range and the AO score positions the answer inside it.
Real mark bands
Marks come from the awarded level's own published mark band, the way an examiner awards them — not back-derived from a flat percentage table.
Caps, gates and floors
Off-topic floors, writing-quality gates and mark-scheme hard caps are applied after scoring — and every one is recorded on the result.
Rate-based accuracy
Technical accuracy is judged on errors per hundred words, so a short paragraph and a long essay are held to the same standard.
Grade transparency
Every grade carries the level it matched, the mark band it came from, and the descriptor wording behind it.
Grade and marks locked — nothing downstream can move them
Feedback Generator
AI call · grade already locked
Only now does the final AI call run, and it is handed a grade it has no power to change. It writes the examiner's explanation — strengths, next-grade steps, AO-by-AO commentary and inline annotations pinned to the student's own sentences.
strengthsimprovementsnextGradeAO commentaryspan annotations
Submission pipeline · validated end-to-end
1Screen
›2Route
›3Extract evidence
›4Judge level
›5Compute grade
›6Award marks
›7Write feedback
›8Deliver
Every AI response is validated against a strict schema, with an automatic retry and a dead-letter queue on malformed output. Each step is idempotent and resumable — a retry can never produce a different grade.
Deterministic
Same evidence in, same grade out. Once stored, a grade is locked and never rewritten.
Replayable
Every result is stamped with the engine and rule-set version that produced it, so any mark can be reproduced later.
Traceable
Each grade cites the level, mark band and descriptor behind it. Nothing is asserted without a source.
Continuously proven
A locked set of fairness invariants — ordinality, determinism, cross-board parity, descriptor traceability, bounded marks — runs on every build.
Board coverage. Rubrics, assessment-objective weightings and level descriptors are transcribed board by board and stored as data — AQA, Edexcel, OCR, WJEC/Eduqas and Cambridge, across English Language and English Literature. Where a board publishes a real level grid, the engine anchors to it directly; where it doesn't, the engine says so rather than inventing one.
Features
Everything you need to level up
A focused toolkit for GCSE English — not a generic AI chatbot.
Core
Instant GCSE grading (1–9)
Real grades in seconds — built around actual exam-board mark schemes and AO weightings.
Learn more
AO1 · AO2 · AO3
AI examiner feedback
Feels like a real examiner, not a chatbot. Specific, on-topic, never vague.
Learn more
Premium
Improving
Improve My Answer
Your answer, rewritten to the next grade — so you can see exactly what to change.
Learn more
Premium
Streak saver
Streak freeze
Protect your revision streak when life gets busy — bank freeze days so one missed day never resets you.
Learn more
Compare
Model answer comparison
See your answer side-by-side with an examiner-level model, with the key differences called out.
Learn more
Tracking
Progress dashboard
Track every grade, every AO, every paper. Watch your average climb week by week.
Learn more
Papers 1 & 2
Language & Literature
Full GCSE Language and Literature support — papers 1 & 2, all major boards.
Learn more
Real papers
Exam practice papers
Sit full past-paper-style exams in a timed environment and get marked section by section.
Learn more
5–10 min
AO drills
Quick 5–10 minute targeted exercises that sharpen one assessment objective at a time.
Learn more
Plan first
Essay planner
Build a structured plan with AI guidance before you write — so every paragraph earns marks.
Learn more
Rewrite
Fix this paragraph
Rewrite a weak paragraph and watch the grade move — the fastest way to learn what works.
Learn more
Literature
Quote flashcards
Master key Literature quotes with spaced-repetition flashcards that resurface what you forget.
Learn more
Gamified
Streaks, badges & challenges
Build a daily revision habit with streaks, weekly challenges and milestones worth chasing.
Learn more
Speed
Timed quick practice
10–15 minute focused tasks under exam pressure. Perfect for revision sprints between lessons.
Learn more
Consistent
Deterministic marking
Same answer, same grade — every time. Anchored in code, not vibes.
Learn more
Reliable
Validated outputs
Every AI response is schema-checked end to end. Bad outputs are caught and retried.
Learn more
Practice
Likely exam questions
Curated, AI-generated questions from common GCSE themes — clearly labelled.
Learn more
Family
Whole family
Parent dashboard
One family plan, every child. Parents follow progress and revision habits in plain English.
Learn more
Family
Email digest
Weekly parent digest
A branded weekly email summarising each child's grades, streak and areas to work on.
Learn more
Product tour
See Marked in action
Real screens from the product — the exam hall, the examiner and your progress, exactly as students see them.
Sit the real paper, virtually
Full past papers with sections, a live timer and locked submissions — the same pressure as the exam hall, from your desk.
Subjects
Two modules. One examiner.
Full coverage of GCSE English Language and Literature — calibrated per exam board and year group.
Module 01 · English Language
Writing that scores.
Imaginative, narrative and transactional writing — marked against the same AOs your examiner uses, with prompts modelled on real GCSE papers.
- Descriptive & narrative writing
- Articles, speeches, letters, reports
- Quick Practice Mode (10–15 min tasks)
- Instant strengths · improvements · next-grade
Module 02 · English Literature
Essays, quotes & unseen poetry.
Practice GCSE essays with AO1, AO2 and AO3 feedback; revise quotes with adaptive flashcards; tackle unseen poems with examiner-style interpretation.
- AO1 / AO2 / AO3 feedback
- Quote memorisation + flashcards
- Unseen poetry interpretation
- Improve My Answer (premium)
Why it's different
Not another AI chatbot. A real AI examiner.
Built specifically for GCSE English — with the consistency of a mark scheme and the warmth of a great tutor.
Generic AI
Marked
Random AI responses that change every time
Deterministic grading — same answer, same grade
Generic 'good job' feedback
Targeted AO1, AO2, AO3 examiner comments
Makes up mark schemes
Built on real GCSE expectations and exam boards
Feels like a chatbot
Feels like a real examiner marking your paper
Hallucinates content you didn't write
Strict guardrails: no invented quotes, no padding
Slow and unreliable
Async queue + auto-retry — feedback in seconds
Consistent by design
The grade comes from deterministic logic. AI is only used to explain it — so your mark never drifts.
Built on AO criteria
Every mark traces back to AO1, AO2 or AO3 — the same language used by real examiners.
Exam-accurate
Question types, mark weights and tone are modelled on actual GCSE English papers.
Student-friendly
No jargon walls. Clear, kind feedback that tells you exactly what to do next.
Loved by students
Real grades. Real progress.
Students across the UK use Marked to push past plateaus and into top GCSE bands.
Honestly feels like a real examiner — not a chatbot. The Improve My Answer feature is unreal. It's become part of my revision routine.
T
Tomas K.
Year 10 · Edexcel
Same answer, same grade, every time. That consistency is what made me trust it. My English grade has gone up by a whole band.
A
Aaliyah R.
Year 11 · OCR
Quote flashcards + unseen poetry feedback are the best part for me. It tells me what I missed without making me feel stupid.
D
Daniel S.
Year 9 · WJEC
Pricing
Simple pricing, serious results
Start free. Upgrade only when you're ready to unlock unlimited marking.
Starter
Try the AI examiner — see how marking works.
£0/ forever
- 5 AI markings per day
- Instant GCSE grade (1–9) with AO feedback
- AO drills, essay planner & timed practice
- Quote flashcards, streaks & challenges
- Exam practice papers & progress dashboard
Most popular
Pro
Unlimited marking and the full examiner toolkit.
£9/ month
- Everything in Free, with no daily caps
- Improve My Answer (AI rewrite)
- Streak freeze
- Priority response speed
Pro+ Family
Two students, one bill — perfect for siblings.
£14/ month
- Everything in Pro
- Up to 2 linked student profiles
- Parent dashboard & weekly digest
- Priority email support
No card required for Free · Cancel anytime · 14-day money-back on Pro
FAQ
Honest answers, no fluff.
Everything students and parents ask before signing up.
Yes — Marked is calibrated against a reference set of marked GCSE answers and lands within ±1 grade of a human examiner ≥85% of the time. Grades come from deterministic logic, not a chatbot, so the same answer always returns the same grade.
Live and ready for studentsStart improving your
Start improving your
grades today.
Write an answer, get graded, and see exactly how to move up. Free to start — no card needed.
✓ GCSE grades 1–9✓ AO-based feedback✓ Works on mobile✓ Cancel anytime