For universities

The grading layer
your university
can stand behind.

UniRubric reads each submission against your rubric and drafts a grade with quoted evidence — embedded in Canvas, Moodle, and Blackboard. Every grade defensible. Every marker still in charge.

Lecturer-in-the-loop · Rubric-anchored · Audit-defensible

A Note from UniRubric

Marking has always been judgement under fatigue. By essay fifty, the third-marker drift sets in. By essay seventy, the rubric you wrote in week one is competing with whatever your eye is tired of looking at. UniRubric does not replace that judgement — it removes the conditions under which it degrades.

Every input is yours. Every output is derivable from your own data. We invent no “burnout reduction” figures or “feedback quality” multipliers — that's the trust mechanism.

ROI calculator

What marking is taking from your staff,in your numbers.

We don't pitch UniRubric as a way to spend less on marking. We pitch it as a way to move marker labour from triage to judgement. The hours below are the ones your staff get back — to spend on the work only a human can do.

20,000 assessments per year · 5,000 students × 2 per term × 2 terms · 200 per staff member
Inputs
Marker labour today
10,000 hrs
20,000 assessments × 30 min · $650,000 loaded labour cost
Marker labour with UniRubric
2,667 hrs
20,000 × 8 min review · $273,333 loaded labour + platform
Teaching load reclaimed
7,333 hrs
Hours your staff get back across the academic year · equivalent to $376,667 of loaded labour

Estimates based on the inputs above. Actual labour reclaimed depends on rubric complexity, marker variance, and submission length. Pilot trial available to validate against your specific cohort. Marking still takes time with UniRubric — about 8 minutes per assessment for review, edit, and approve. What changes is what those minutes feel like at hour six of marking.

See the labour load against your real cohort. The full configurator is on the next page.

Configure your contract →
Why universities choose us

Four reasons,
none of them
cost-cutting.

Universities do not buy UniRubric to spend less on marking. They buy it to mark fairly, retain teaching staff, and stand behind every grade.

  1. 01

    Fair marking, every time

    Every grade is anchored to your rubric with verbatim quotes from the criterion descriptor and the student's submission. Marker fatigue, drift, and bias don't decide outcomes — your rubric does.

  2. 02

    Markers who don't burn out

    The cognitive load of marking eighty papers at week twelve is what drives sessional staff away. Take the load off, keep the judgement, keep your people.

  3. 03

    A reputation for consistency

    When the third essay in the pile is judged by the same standard as the eightieth, students notice. So do appeals committees, accreditation reviewers, and prospective enrolments.

  4. 04

    Audit-defensible by design

    Every grade carries the rubric phrase it scored against and the evidence phrase from the student's work. A full audit trail is the default, not a feature you have to enable.

Trust posture

The standards procurement, IT, and academic-integrity offices ask about — the ones we lead with, not the ones we hide behind a sales call.

  • Sydney region

    All student data and rubrics stay in Australia. Region-pinned to ap-southeast-2.

  • AU Privacy Act

    APP-aligned data handling. Australian Privacy Principles applied across the platform.

  • GDPR & FERPA

    European and US student-data frameworks honoured for international cohorts.

  • Your data, never trained on

    Submissions and rubrics are never used to train our models. Yours and yours alone.

  • Lecturer-in-the-loop

    No grade reaches a student until your lecturer reviews, edits, and approves.

  • Full audit trail

    Every grade carries the rubric phrase and the evidence phrase. Defensible at appeal.

See it work on a real assignment

The institutional sandbox is free, runs against your own rubric, and ships results in a week.

Or sign up directly and run a single marking job free, today.

In the systems you already use

No second tab.
No copy-paste.
No setup tax.

UniRubric runs inside Canvas, Moodle, and Blackboard. When a student submits, a draft grade is waiting in your LMS by the time you open it. You review, edit, approve, release — without ever leaving the grading screen you already know.

We are building the institutional product to disappear into your workflow. The friction other AI marking tools introduce — the new logins, the export-import dance, the parallel grade book — is exactly what we are here to remove.

A note from the founder
“The first AI grading tools your faculty trialled were not bad at marking. They were bad at fitting into the day. New logins, export-import dances, parallel grade books — and within a week the cohort had quietly returned to manual. UniRubric is the answer to that. Embedded in the LMS your faculty already use. Lecturer in charge of every grade. Every score defensible at appeal.”
Nawras Alali · Founder & Chairman, UniRubric · Adelaide
See it work

The institutional product,
embedded in the LMS your faculty already use.

A scripted walkthrough of one marker's morning — PSYC2040, Adelaide College of Psychology, 64 second-year students, auto-graded on submit, reviewed and released without leaving the LMS.

Adelaide College of Psychology|Semester 1, 2026
PSYC2040AssessmentsCommunication Skills Reflection
UniRubricLTI 1.3 · Connected
Embedded · auto-launched
PSYC2040 · Marking job
Communication Skills Reflection
·reid-lachlan-reflection.docx
278 words

Communication Skills Reflection

During my placement at the after-school program I had a long conversation with Jay, age nine, about why he had hit another child during the day. My early questions were closed — "did you hit him?" — and his answers were short and defensive. The conversation only opened up when I asked him to tell me what had happened from the start.

Jay is at the concrete-operational stage (Piaget, 1972), which means he can reason about specific events but still struggles to hold competing perspectives at once. My closed questions had asked him to confirm a verdict; my open question asked him to lay out a sequence. The second was developmentally more accessible.

What I learned was that I had been treating Jay as if he should be able to take responsibility in the abstract, when what he actually needed was to retell the event so we could reason through it together. Once he had narrated the chain of small frustrations that led up to the hit, he reached the conclusion that he could have walked away — without me telling him so.

If I were to do this again I would start with the open question. The closed questions cost us the first few minutes of the conversation and made Jay feel accused before he was given the chance to explain. The reflection has made me more cautious about questioning techniques that assume cognitive moves a child has not yet made.

The course concept I had not appreciated before this placement was that question structure is itself a developmental scaffold. The shape of the question quietly tells the child what kind of thinking we expect of them.

Overall
76/ 100Distinction

A solid reflection that gets stronger as it progresses. The shift from closed to open questioning is well-observed, and the closing line about question-shape as scaffolding is a real insight. Tighten the middle and this is a high-distinction piece.

Student submission
reid-lachlan-reflection.docx
278 words · submitted via LMS
PSYC2040 · Reflection

Communication Skills Reflection

by Lachlan ReidSubmission ID · LACHLAN

During my placement at the after-school program I had a long conversation with Jay, age nine, about why he had hit another child during the day. My early questions were closed — "did you hit him?" — and his answers were short and defensive. The conversation only opened up when I asked him to tell me what had happened from the start.

Jay is at the concrete-operational stage (Piaget, 1972), which means he can reason about specific events but still struggles to hold competing perspectives at once. My closed questions had asked him to confirm a verdict; my open question asked him to lay out a sequence. The second was developmentally more accessible.

What I learned was that I had been treating Jay as if he should be able to take responsibility in the abstract, when what he actually needed was to retell the event so we could reason through it together. Once he had narrated the chain of small frustrations that led up to the hit, he reached the conclusion that he could have walked away — without me telling him so.

If I were to do this again I would start with the open question. The closed questions cost us the first few minutes of the conversation and made Jay feel accused before he was given the chance to explain. The reflection has made me more cautious about questioning techniques that assume cognitive moves a child has not yet made.

The course concept I had not appreciated before this placement was that question structure is itself a developmental scaffold. The shape of the question quietly tells the child what kind of thinking we expect of them.

End of submission · 278 words
0:001 / 121:22
Responsible adoption

An assistant to your markers — never a replacement.

Every score is a recommendation that your lecturer reviews, edits, and approves before a student ever sees it. Every recommendation links to a verbatim phrase from your rubric and a verbatim phrase from the student's submission. Every action is captured in a full audit trail.

  • Lecturer-in-the-loop on every grade. No auto-released marks.
  • Every score links to a quoted rubric phrase and a quoted evidence phrase from the student's work.
  • Your data never trains our models. Region-pinned data residency. AU Privacy Act, GDPR, and FERPA aligned.
UniRubric — The grading layer your university can stand behind.