Feature

A question bank that refuses to store the same question twice

The question bank is a per-store library that questions are drawn from rather than retyped into each test. Its job is to stop the same question existing in five slightly different spellings.

Duplicate detection

Every question is fingerprinted on its normalised content. An exact duplicate is caught outright; a near-duplicate — the same question with different punctuation, whitespace or a reworded stem — is caught by trigram similarity and flagged for you to confirm, merge or keep.

The same pipeline runs on bulk imports, so uploading a spreadsheet that overlaps your existing bank links the existing questions instead of creating a second copy of each.

Versions and translation families

Immutable version history: Editing a question’s content appends a new version rather than overwriting. A report from six months ago still reproduces the question as it was when it was answered.
Translation families: The same question in several languages is grouped into one family, so a bilingual test draws one question and delivers it in the candidate’s language.
Usage tracking: Each question records where it has been used, so you can see whether a question has already been shown to a cohort before adding it to a new paper.
Status and lifecycle: Questions move through draft, published and archived, and the console filters on status, duplicate status, tags, difficulty and language.

Getting questions in and out

Upload a spreadsheet; it is parsed and validated in the background.
Review the preview — per-row summary, errors, warnings, and which rows matched existing bank questions.
Fix rows inline, or import only the valid ones and re-upload the rest.
Export the bank whenever you need a copy outside the platform.

What this does not do

Worth knowing before you plan around it.

The bank is scoped to a store. Separate stores keep separate banks, which is deliberate — it is what stops one client’s paper leaking into another’s. There is no cross-store shared bank today.
Near-duplicate detection is a similarity signal, not a judgement. It surfaces candidates for review; you decide whether two questions are genuinely the same.

Common questions

What happens if I edit a question that has already been answered?

The edit creates a new version. Attempts that were graded against the earlier version still reference it, so historical reports reproduce exactly what the candidate saw.

Can I import questions from a spreadsheet?

Yes. Uploads are parsed and validated in the background and you get a preview with per-row errors and warnings before anything is committed. Rows that match questions already in the bank are linked rather than duplicated.

Stop retyping questions

Import what you already have, and let the bank catch the duplicates.

See plans and pricing Assessments for coaching institutes