Grading at Course Scale, Marked to Your Criteria
Two hundred submissions, a grader or two who each mark slightly differently, and a semester that does not slow down for any of it. GradingPal marks against the criteria you wrote, keeps the reasoning and the student’s own words behind every score, and gives everyone marking your course the same starting point.
Start free. See what each plan includes for course-sized volume.
The Part of the Job Nobody Costed
Contact hours are scheduled, office hours are posted, and the forty hours of marking between them are assumed. If you are teaching four sections across two institutions on adjunct contracts, they are assumed twice. Nobody in higher education has ever been given more time for grading; the only variable anyone can actually change is how long a defensible mark takes to produce.
If You’re on Canvas, Read This First
GradingPal does not integrate with Canvas. You can still use it — download the submissions, grade them here, and take the results back — and plenty of instructors do exactly that. But if a one-click integration is what you need, we are not that yet, and you should know it now rather than after you have set up a course. The manual route, step by step is written out in full, along with where Canvas sits on our list.
Nobody Sets Your Criteria But You
There is no state framework for a 300-level seminar. The criteria are yours, they are on your syllabus, and they are the thing a student will hold you to. Bring your rubric as written — the wording, the weightings, the levels — and the marking follows it criterion by criterion, showing you the reasoning and the passage it is judging. Add the conventions you never wrote down (what a citation has to include, whether late work is capped) and they apply to every assignment in the course.
Two Graders, Two Standards, One Complaint Email
A course with graders has a problem no single-teacher classroom has: students in section 4 compare marks with students in section 2, and by week six somebody has written to the department about it. Share the course and everyone marking it works from the same rubric, the same written conventions, and the same evidence behind every suggested score — it is one course, several markers. They keep their own judgement — that is the job — but they no longer start from four different readings of what a B is.
It does not make graders agree. It makes disagreement visible, on the criterion where it happened.
Feedback Written to an Adult
Your students are adults paying for the course, and the comment that lands with a fourteen-year-old is not the one that lands with them. Feedback is written to the student, at the level of the work, ending with the specific next step — and you choose the register: direct and rigorous for a graduate seminar, warmer for a first-year survey, questions rather than answers where you want them to work it out. What the comment sounds like is set once per assignment, not negotiated per student.
Everything a Course Sets
- Papers and written assignments
- The bulk of it, marked criterion by criterion with the passage quoted. papers and written assignments.
- Problem sets and lab reports
- Method marks, units, whether the conclusion follows from the data. lab reports and calculations.
- Midterms and finals
- Including blue books and handwritten scripts, scanned as one file. handwritten scripts.
- Presentations, capstones and vivas
- The deck and the talk, marked separately. recorded work.
- Discussion posts and short responses
- Where the volume is weekly and the marking is a criterion or two each.
What It Can’t Do
It won’t integrate with Canvas
Not yet. The manual route is documented, and you should decide on that basis rather than on a roadmap.
It won’t make your graders agree
A shared rubric is a starting point. What changes is that disagreement becomes visible on the criterion where it happened.
It won’t detect AI-written work
By choice. The full reasoning sits on the responsible-AI page, and it is a permanent no.
It won’t set the final grade
Every score is a suggestion until you approve it, and nothing reaches a student until you release it.
More on what we will not claim to detect.
Try It on One Assignment
Start on the free plan and run one assignment through — a set you have already marked is the fastest way to see whether the judgements hold. Course-sized volume needs a paid plan; see what each plan includes.
Frequently Asked Questions
My whole department runs on Canvas. Is this usable?
Not as an integration. You download submissions from Canvas, grade them in GradingPal, and take the results back — a route we document step by step. If a one-click integration is a requirement for you, we are not there yet, and Canvas is on our list.
Can my TAs and I all grade the same course?
Yes. Share the course and everyone marking it works from the same rubric, the same written conventions and the same evidence behind each suggested score. They keep their own judgement — what changes is that they start from the same place rather than from four readings of what a B is.
Can it grade handwritten exam booklets?
Yes. Scan a set as one file and it separates into a submission per student, reads the handwriting and marks each against your criteria. Where something genuinely cannot be read, it flags that question for you instead of guessing.
Will it grade to my course learning outcomes?
It grades to the rubric you bring, in your wording, criterion by criterion — which is where course outcomes usually live.
Is it worth it for a small seminar?
The saving scales with the set, so a twelve-person seminar is not where it earns its place — a two-hundred-student survey with graders is. What holds at any size is that the reasoning behind every mark is kept with the work.
Run One Assignment Through
Bring a set you have already marked and your own criteria. The point is not speed — it is that the reasoning behind every mark stays with the work.