FREQUENTLY ASKED
What is Lectora?
Lectora is an AI grading assistant for university educators. It drafts grades and feedback on long-form exam answers — clinical cases, financial analyses, mathematical proofs — against your rubric. The educator reviews, edits, and signs off. Built by Fjordbyte in Norway and validated against 895 candidates on a real medical exam, where the AI's draft agreed with the course teacher (R² = 0.81) more closely than two independent human graders agreed with each other (R² = 0.64).
What is Lectora, in a sentence?
Lectora drafts the grades and feedback for an entire cohort, the educator approves them. The educator's name is on the result; Lectora's name isn't.
That's the design constraint. Lectora is not an autograder, not an assessment-replacement product, not an attempt to take the educator out of the loop. The grading workflow stays human — Lectora makes the human's part of it dramatically faster by handling the first-pass draft, the cohort-wide pattern analysis, and the per-student feedback that's the slowest part of long-form grading.
Who builds Lectora?
Fjordbyte AS, a Norwegian software company. Lectora is Fjordbyte's flagship product. The team is small, technical, based in Norway, and built the validation methodology in partnership with the University of Bergen's medical faculty. Pricing, support, and contracting are all handled directly — there's no reseller layer.
How does data privacy work?
Lectora is GDPR-compliant. Student data stays inside the EU, processed by EU-hosted infrastructure. Student answers are not used to train any AI model — not Lectora's, not any underlying provider's. Data is encrypted in transit and at rest, retention is configurable per institution, and we sign DPAs as a standard part of onboarding. The cookies and tracking inventory is published at /cookies for both locales.
How is Lectora priced?
Per-course or per-institution, with pilot pricing for first deployments. No public price list — pricing depends on cohort size, number of courses per semester, and contract length. There's no free tier today and no per-student fee. Quotes are issued after a demo and an initial scoping conversation.
How does Lectora fit into our existing grading workflow?
Out of the box: Lectora installs as an LTI 1.3 app in your Canvas instance. Instructors launch it from inside the course they already work in, on the assignments they already grade, against the rubric they already use — no new portal, no new login, no parallel grading database to keep in sync. For institutions evaluating before a full install, instructors can connect a Canvas Personal Access Token and run a pilot without IT involvement; the data-flow is identical (see /canvas-access for the full PAT mode description).
Beyond out of the box: every institution grades a little differently — rubric format, sub-question structure, conventions for partial credit, how feedback is phrased and where it's posted. The product as it ships covers the common shape; fitting it to your specific workflow is what the pilot semester is for. We work directly with the course team during onboarding to extend rubric coverage, adjust feedback templates, and wire Lectora into the parts of the workflow that touch other systems — exam-platform exports, archiving, student records — within the technical limits of those systems. What we won't do is promise integration with systems we can't responsibly touch. We're explicit about those limits upfront, not after the contract.