Feature

Online assessments with server-enforced timing

The assessment engine covers authoring, delivery and grading. Timing is enforced server-side, not by the candidate’s clock, so an attempt cannot be extended by pausing a laptop or changing a system time.

What you can author

A test is made of sections, and each section holds questions in one of four types. Objective questions grade automatically the moment an attempt is submitted; short-answer questions can be routed through a manual-review lane before a score is final.

Four question types: Single choice, multiple choice, true/false, and short answer. Each question carries its own marks and optional negative marking.
Sections with their own rules: Order, marks, and (in section-wise mode) a duration per section, so a quantitative section can be tighter than a reading section.
Rich text and images: Question prompts and options are rich text, so formulae, code blocks and diagrams survive into the candidate runtime and the report.
Multi-language delivery: A question can carry translations. When language switching is enabled on a test, the candidate picks their language and the report is generated in the language they actually sat the test in.

How timing works

Two timing models are available, and the choice is structural rather than cosmetic. `ENTIRE_TEST` gives one deadline for the whole attempt; `SECTION_WISE` gives each section its own deadline and the total is their sum.

Both are anchored to the server. The candidate runtime displays a countdown, but the deadline that matters is computed from the attempt’s server-recorded start time — so a lost connection, a refresh, or a clock change does not buy extra time. On expiry the attempt (or the section) auto-submits whatever has been answered.

Delivery and navigation

Navigation policy: Per test: allow section jumping, allow returning to previous questions, allow moving forward before a section is complete, auto-advance when a section expires, and whether an expired section stays readable.
Randomisation: Question and option order can be randomised per candidate. The order each candidate saw is snapshotted, so their report and the grading match what was on screen.
Attempts and passing score: Maximum attempts and a passing score are set per test; both are enforced when an attempt starts and when it is graded.
Autosave and resume: Answers save as the candidate works. A dropped connection resumes the same attempt against the same server deadline.

After the attempt

Every submitted attempt automatically emails the candidate an exam report: their own answers, in the language they took the test in, with no correctness or scoring information. It is a reference copy, not a result.

The scored breakdown — correct answers, marks, and pass/fail — is a separate email that is only sent when you explicitly share it from the results view. Analytics break results down by section, by question, and by time-on-task.

What this does not do

Worth knowing before you plan around it.

Once the first candidate starts an attempt, the timing model and section durations are locked for that test. Other fields stay editable; changing the timing model means duplicating the test into a fresh draft — which is what keeps in-flight and historical reports honest.
Short-answer questions are graded by a person unless you accept an exact-match rule. Plan review time into cohorts that use them.

Common questions

Can a candidate get extra time by disconnecting?

No. The deadline is derived from the server-recorded start time of the attempt, not from a client timer. Reconnecting resumes the same attempt against the same deadline, and expiry auto-submits whatever has been answered.

Do candidates need an account to sit a test?

Usually not. Most assessments accept candidate registration by email. Private or paid tests can require sign-in, a subscription, or a purchase first, and the portal walks the candidate through whichever applies.

Can one test run in more than one language?

Yes. Questions can carry translations, and enabling language switching on the test lets each candidate choose. Their report is generated in the language they sat the test in.

Build your first test

Author a test, invite a cohort, and read the analytics on one plan.

See plans and pricing Open the assessment portal