GRADING
How a batch is graded
One run drafts a score and a reason for every submission.
Read the page →
FEATURES
What can Lectora actually do?
Lectora is a trusted AI workspace for course assessment in higher education. It runs inside Canvas via LTI 1.3, works from the material and learning goals your course has declared, and hands every score, comment and question to a person as a proposal. This page is the index: one page per feature, each stating what the feature does and what it does not.
One page per feature
FEEDBACK
What the student receives
Written feedback for every submission, not only the borderline ones.
Read the page →
COHORT ANALYTICS
Reading the cohort
Internal score distribution and summary statistics for one assignment.
Read the page →
CURRICULUM
What Lectora reads before it drafts
Indexed files, passages and learning goals. Nothing is applied until you accept it.
Read the page →
QUIZ STUDIO
Drafting exam questions
Multiple choice, short answer and extended response from the indexed material.
Read the page →
COURSE ASSISTANT
Asking your course material
Ask the indexed material and open the passage the answer came from.
Read the page →
DOCUMENTS
Document understanding
Scans, documents and spreadsheets as the basis for a draft you review.
Read the page →
The limits, in one place
Lectora integrates with Canvas and no other LMS. The in-course assistant can read Canvas data but cannot change it; what Lectora writes there (grades, grading comments, classic-quiz scoring) happens as an educator's own act, under that educator's Canvas identity, and is published operation by operation at /en/canvas-access.
We hold no SOC 2 and no ISO 27001, we are not listed in 1EdTech's certified-product directory, and we have appointed no Data Protection Officer. We say so here because a serious procurement reader establishes it in the first meeting either way.
On quality we claim what has been measured and nothing further. Score agreement on long-form written exams is measured and published. Question quality is not, and the score-agreement result does not transfer to it.
It drafts scores and reasons for submissions, drafts written feedback, and shows an assignment's internal score distribution. Curriculum connects indexed course material, learning goals and assessments. Quiz Studio drafts questions for educator review. The course assistant lets you ask about the indexed material. Document understanding explains how scans, documents and spreadsheets support this assessment work.
No. Every output is a proposal. For a grade, the person who acts on it must be a qualified examiner; that is Norwegian law, not a product setting. The Universities and Colleges Act of 8 March 2024 no. 9 reserves grade-setting for qualified examiners, and examiner approval is the human-oversight measure Lectora has declared under Article 14 of the EU AI Act.
Inside your Canvas course, installed as an LTI 1.3 tool, reached through the same login and the same roster. There is no second account to provision and no parallel tool surface. Processing runs on European infrastructure under a published Data Processing Agreement.
Canvas, via LTI 1.3. That is the only integration that ships today. Anything else you may have seen discussed is not available, and this page will say so when that changes.
The published validation covers score agreement on long-form written exams: twelve MED12 sittings at the University of Bergen, 889 candidate-exam pairs. It does not extend to feedback quality, question quality, coverage usefulness or retrieval quality, and we do not stretch it to cover them. The same validation pages carry the independent study where AI assistance made an uncalibrated grader pool less consistent.
The feature pages cover grading, feedback, cohort analytics, curriculum, Quiz Studio, the course assistant and document understanding. Each states what is supported and, where it matters, what is not.