AI-assisted grading for handwritten answers.

Overview

Students write subjective answers on plain paper as usual. Upload the scans — local OCR reads the handwriting (math/LaTeX-aware) and AI grades against your rubric, with teacher override on every score.

A board-style written exam, set up like an assignment

A guided wizard walks you from title to publish: link any Paper Canva paper as the question source, set the exam date, due date and submission window, write the rubric with marks against each criterion, and choose how much of the grading the AI should do — fully automatic, AI with teacher review, or entirely manual. That last choice is a setting rather than a philosophy you have to accept from the vendor. Submission policy is yours too: which file types, what size limit, whether resubmissions are allowed, what a late paper costs, and whether marking happens anonymously.

AI proposes the marks; you keep the red pen

Students upload photographs or scans of their handwritten answers — a phone picture of a notebook page is enough. OCR reads the writing, mathematics and LaTeX included, and the AI scores each answer against your rubric with its reasoning attached, so a proposed mark can be agreed with or argued down rather than merely accepted. You then work in a review workspace built for the job: annotate directly on the submitted PDF, move question by question with keyboard shortcuts, and adjust anything that looks wrong. Every score is overridable and the teacher’s mark always wins. Nothing reaches a student until you release it.

Subjective marks that flow like objective ones

Descriptive papers usually end their life as a number in a register, which is why nobody ever learns anything from them twice. Released marks here join the same analytics as your online and OMR exams, rolled up by board, class and subject and broken down per student and per chapter. Students see the annotated sheet itself, the rubric breakdown that produced their mark, and your overall feedback — the moment you publish, and not before. AI insights on each submission help you write sharper comments than the third "needs more detail" of the evening, and the whole set exports when the office asks for it.

Term-exam correction that finishes on time

Every answer script is scanned once and graded in one workspace. AI does the first pass against the rubric; teachers review, annotate and release — and the marks are already in analytics when report cards come around.

Board-pattern descriptive tests, weekly

Homework that gets real feedback

Students photograph their notebook pages and upload. You review AI-proposed marks over tea, scribble on the PDF where it matters, and every student gets annotated, rubric-based feedback — not just a tick.

Exam setup

Evaluation modes

Rubrics & marking

Student submission

Grading workspace

Results & analytics