Lose the grading pileup.

CompGrader turns one scan of a class's handwritten work into organized, reviewable grades and individualized feedback for each student. In any subject.

Try the live demo Learn more

CompGrader class workspace with multiple classes, assignments, progress metrics, and a roster

Made for the work in between

Not every assignment needs human grading

Use CompGrader for the frequent, handwritten work where completion matters: practice, early thinking, and evidence that students engaged with the task.

Daily practice

Check whether students attempted the expected work without turning every practice set into a correctness grade.

Starters and sketches

Move through bell-ringers, brainstorming pages, diagrams, and other handwritten thinking more efficiently.

Let the computer do the sorting

For larger classes, organizing student submissions alphabetically can be a pain. Let the AI grader do that for you.

From copier to gradebook

The tedium goes. Your judgment stays.

A simple workflow handles the mechanical steps while leaving the consequential ones with you.

01
Bring the roster
Create a class and paste in the student names you already keep in your LMS or gradebook.
02
Scan the stack
Feed the papers through the copier and upload the result as one PDF - no splitting or renaming files.
03
Review the suggestions
CompGrader matches students with their submissions, provides completion estimates, and suggests grades with feedback.
04
Approve and export
Make any corrections, preview the final grade rows, and download a clean CSV for import into your LMS.

Inside the workflow

See what changes at each handoff

The intake

One upload

Choose the class, upload the scan, and give grading instructions.

One PDF for the whole stack
Assignment structure you define
Roster context included

CompGrader new assignment form with a scanned PDF and completion grading settings

The review

Put instructor attention where it matters

See missing, ambiguous, or low-confidence work first. Original PDF page numbers stay visible so every suggestion can be traced back to the scan.

Exceptions surfaced clearly
Source pages kept in view
Page matches remain editable

CompGrader results overview with student scores, source pages, and varied completion statuses

The handoff

Leave with grades you have approved

Adjust completion status, percentages, scores, and export inclusion before anything leaves the app. Then preview the rows and download your CSV.

Editable completion results
Final grade preview
Spreadsheet-ready CSV

CompGrader downloads and exports screen with original, sorted, and per-student files

Choice and privacy

Maximizing an instructor's academic freedom and student privacy.

We want to expand choice, not restrict it. We let the instructor decide which underlying AI model is the engine for the grading, including open weight models from Microsoft, Google, Alibaba, or one you design.
With swappable open-weight models, student data does not need to be sent to large technology companies.
Universities can self-host CompGrader on their own servers.
CompGrader does not train on student data.
We plan to redirect any profit to scholarships for the types of students who were traditionally hired as grading assistants.

Interested in trying this out?

We're testing whether enough educators want it before we procure all the resources to run the full product.

A token deposit helps us validate demand and decide what to build next

Put down a token deposit today Explore the live demo