Built on the IB’s published assessment criteria

More than a mark.
A report you can write from.

essaycriteria sends a swarm of specialist AI examiners through your IB DP essay. Every comment is pinned to the official descriptor it belongs to, every band decision quotes your own writing — and a next-draft roadmap tells you what to fix first.

  • EE · TOK · Group 1–6
  • ~4-minute turnaround
  • Every claim quotes your text
Rubric Keeper
Evidence Hunter
Argument Architect
Language Examiner
Devil’s Advocate
The Chair
SWARM ANALYSISEE · History · draft 2

    Live view — a swarm marking an EE in History (the Berlin Airlift).

    6specialist agents read every essay
    100%of comments cite a descriptor & quote your text
    ~4 minfrom submission to full report
    2,000+words of targeted feedback per essay
    The problem

    A number can’t teach you to write.

    Most AI graders compress four thousand words of thinking into a single number. “27/34” tells you where you landed — never why, and never what to change.

    A typical AI grade
    27/34
    • Which criterion?
    • Which descriptor missed?
    • Which paragraph?
    • What to change?

    Four thousand words in. One number out.

    The essaycriteria report
    • Criterion A · 5/6 — “RQ sharply scoped; methodology implicit — name your archives.”
    • Criterion C · 9/12 — “Evaluation appears in §3, then vanishes exactly where it matters.”
    • The one move that buys the next band — and the paragraph to make it in.

    Every line pinned to a descriptor. Every claim quoting your text.

    The swarm

    One AI grades. Six read.

    An IB essay isn’t one problem — it’s at least five, and they hide inside each other. So we don’t send a generalist. We send a team of specialists, and a chair that makes them argue.

    Rubric Keeper

    Holds the official descriptor set for your exact task. Nothing enters the report unless it names the criterion — and the descriptor — it answers to.

    Criterion pinsBand discipline

    Evidence Hunter

    Reads every claim against its support. Finds the sentence where analysis quietly slipped into description.

    Claim→evidence mapDescription drift

    Argument Architect

    Sketches the skeleton of your argument — thesis, development, synthesis — and finds the load-bearing sentences that aren’t there.

    Structure mapCounter-claim audit

    Language Examiner

    Checks register, precision and subject terminology against the language of your discipline.

    Register scanTerminology

    Devil’s Advocate

    Attacks your argument the way a rigorous examiner would — then checks whether your essay already has an answer.

    Stress testsRebuttal check

    The Chair

    A chief-examiner agent. Cross-examines the other five: no critique survives without a quote from your essay and a descriptor reference. Then signs the report.

    Consensus rulingFinal verdict

    Agents don’t just report — they argue. When the Evidence Hunter calls a claim “unsupported”, the Devil’s Advocate checks whether the essay answers elsewhere — and the Chair strikes any finding that can’t point to a line number. What survives is insight, not opinion.

    How it works

    From submission to insight in four moves.

    1

    Submit your essay

    Paste the text or upload the PDF. Full 4,000-word EEs, HL essays and RPPF reflections all welcome.

    2

    Choose the exact task

    EE in History. TOK essay on title 3. English A HL essay. The correct official descriptor set loads itself.

    3

    The swarm reads

    Six specialists analyze in parallel, then cross-examine each other. The Chair strikes anything that can’t cite your text.

    4

    Read, fix, resubmit

    Criterion bands with justifications, line-level annotations, chief-examiner verdict, next-draft roadmap — in about four minutes.

    Rubric fidelity

    Pinned to the descriptors — not to vibes.

    The IB doesn’t reward “good writing” in the abstract; it rewards criteria — bands written in cold, specific language. essaycriteria feedback speaks that language natively: every comment carries its descriptor, every decision quotes the band your work meets, and the one it doesn’t. Yet.

    All subject groups · marked out of 34 · Sample student · EE (History) · 23/34 · indicative grade B
    C

    Critical thinking

    Research, analysis and evaluation of evidence.

    7/12
    10–12
    Excellentthe next band

    Argument developed precisely and critically; competing evidence and interpretations weighed; source reliability evaluated and woven into the analysis.

    7–9
    Goodyou are here

    Argument developed with some critical evaluation; reliability considered in places, unevenly integrated.

    4–6
    Satisfactory

    Analysis attempted but uneven; description competes with argument; sources reported rather than weighed.

    1–3
    Insufficient

    Largely narrative or descriptive; little or no evaluation of evidence.

    The gap to 10–12: The 10–12 descriptor pays for evaluation woven in: interpret the tonnage against Soviet expectations (§2), weigh Murphy (§3), and rebut the “division settled by 1947” reading (§4).

    Descriptor text paraphrased for this preview — full official wording is used in reports. essaycriteria is not affiliated with or endorsed by the International Baccalaureate Organization.

    A real-style excerpt

    See the depth for yourself.

    A report on a History EE — the Berlin Airlift, draft 2 of 3, 23/34. Click a number in the essay, a criterion bar, or a legend chip to explore what the swarm found.

    Extended EssayHistoryDraft 2 of 3 · 3,986 words

    RQTo what extent was the 1948–49 Berlin Airlift a turning point in the early Cold War?

    Filter:
    §1 · Introduction

    The Berlin Airlift of 1948–49 is remembered as the moment containment1 became an operational reality rather than a doctrine on paper. This essay argues that the Airlift was a turning point less for what it delivered than for what it ruled out: a military solution to the Berlin question2.

    §2 · Operation Vittles

    Truman faced enormous pressure from his Joint Chiefs to abandon the city3. Over 462 days, the operation delivered 2.3 million tons of supplies at the cost of 101 lives4 — a feat Tunner called “the most consequential logistics operation of the early Cold War.”

    §3 · The view from the ground

    Murphy notes that morale never broke5: “We flew, they stayed, and the city held.” That endurance became, in itself, an argument for the strategy of patience6.

    §4 · Conclusion

    The Airlift was therefore a turning point, because the West learned resolve7. This was very significant8 for Allied cohesion.

    — excerpt continues · §5–§9 · 3,742 words omitted —

    The verdict at a glance

    23/34

    indicative grade B · boundaries vary by session

    C

    Critical thinking

    7/12

    The decisive criterion. Analysis flashes in §2 — then load-bearing claims go uncited, the key witness goes unweighed, and the counter-interpretation goes unanswered.

    The gap: The 10–12 descriptor pays for evaluation woven in — interpret the tonnage (§2), weigh Murphy (§3), rebut “settled by 1947” (§4).
    3

    Load-bearing claim, no citation

    “Enormous pressure from his Joint Chiefs” carries your §2 argument. Which chiefs? When? The JCS memorandum of 28 June 1948 exists — cite it, or cut the sentence.

    pinned to C · 10–12
    4

    Data without interpretation

    462 days, 2.3 million tons — excellent data, still description. The top band pays when you set it against Soviet expectations: was Stalin counting on Western fatigue?

    pinned to C · 10–12
    5

    Witness unweighed

    You cite Murphy, never weigh him: he flew the route himself. Vivid testimony — compromised proximity. One paragraph evaluating him is exactly what the 10–12 descriptor rewards.

    pinned to C · 10–12
    7

    Assertion where rebuttal belongs

    “Because the West learned resolve” — the counter-reading (Germany’s division effectively settled by 1947) is never introduced, never answered.

    pinned to C · 10–12

    The Chair’s verdict

    The work of a historian-in-training who can already build a thesis — this research question is sharper than most supervisors see in a year. The ceiling is Criterion C: analysis flashes in §2, but the essay still trusts its sources and asserts its turning point. Two habits separate this 23/34 from a comfortable A — cite the claims that carry weight, and weigh the witnesses you cite. Both are learnable in one draft.

    — The Chair · synthesis agent

    Next-draft roadmap

    1. Weigh Murphy — one tight paragraph on his proximity to events and the limits it sets on his testimony.

      C → 10–12
    2. Source the JCS claim (the 28 June 1948 memorandum) — or cut it.

      C
    3. Rebut the “division settled by 1947” interpretation before your conclusion.

      C
    4. Rebuild RPPF reflection 2 around one decision and what it changed.

      E → 5–6
    Beyond the score

    Everything a number can’t say.

    Line-level annotations

    No “work on your analysis.” Every comment quotes the exact sentence it’s about — so you know what earned it, and what it costs you.

    Band-boundary coaching

    The IB pays in bands. See the exact descriptor separating your band from the next — and the paragraph where you can earn it.

    Next-draft roadmap

    Findings ranked by marks-per-effort, so revision starts where it pays — not where it’s loudest.

    Chief-examiner verdict

    The Chair synthesizes six agents into one honest, humane paragraph — the kind of synthesis top candidates use to steer their next draft.

    RPPF & reflection coaching

    Criterion E is graded on your reflections. The swarm reads the RPPF too — and coaches the evaluative move each one is missing.

    Draft-over-draft tracking

    Resubmit and watch each criterion move. Bands that didn’t move come with an explanation.

    Pricing

    Start free. Upgrade when the depth sells itself.

    Try one criterion
    $0 / forever

    Feel the depth before you pay a cent.

    • Any single criterion, full swarm treatment
    • Line-level annotations for that criterion
    • Integrity & citation scan
    • No card required
    Start free
    Most chosenFull report
    $12 / essay

    The whole swarm, the whole essay.

    • All criteria — full descriptor mapping
    • Line-level annotated essay
    • Chief-examiner verdict + roadmap
    • RPPF reflection check (EE)
    • Draft-over-draft tracking
    Mark my essay
    Schools & cohorts
    Custom

    Bring the swarm to your department.

    • Teacher dashboard & cohort analytics
    • Bulk submissions & RPPF screening
    • Whole-cohort band-gap heatmaps
    • SSO, invoicing & data-residency options
    Talk to us
    FAQ

    Fair questions.

    Is essaycriteria affiliated with the IB?

    No. essaycriteria is an independent study tool. We mark against the IB’s published assessment criteria because they are public and admirably specific — but we are not affiliated with, endorsed by, or connected to the International Baccalaureate Organization. Your official grades come only from your school and the IB’s examiners.

    How can I trust an AI’s judgment?

    Because our reports are built to be checked, not believed. Every claim the swarm makes must survive cross-examination by the other agents — and must cite two things: the line in your essay it refers to, and the descriptor it is measured against. If a comment doesn’t quote your writing, it doesn’t ship. Disagree with a ruling? The report shows its evidence — appeal to your teacher with it.

    Which tasks are supported?

    The Extended Essay in all subject groups (including an RPPF check), TOK essays on the current prescribed titles, and Group 1 Higher Level essays. Subject-specific task sheets — History Paper 2/3, Psychology ERQs and more — are added continuously.

    What happens to my essay after marking?

    Your writing stays yours. Essays are encrypted in transit and at rest, deleted automatically after 30 days, and never used to train models. Teachers can request deletion at any time.

    Is getting AI feedback okay under academic integrity?

    Yes — if the writing is yours. Our feedback works like a very thorough teacher’s comments: it diagnoses, it never ghost-writes. The integrity scan also surfaces citation anomalies and sudden register shifts, so you can fix them before your supervisor does.

    How long does it take?

    About four minutes for a full 4,000-word EE. The swarm runs in parallel — the Chair is the slow one, reading everything twice.

    Your move

    Stop guessing what the examiner saw.

    One criterion is free. Depth does the selling.