Exam Center
Book a demo
Platform

Objective marks instantly. Long answers with a rationale.

Three scoring paths that converge on one result — and a human confirmation step wherever it matters.

One submitted paper through three scoring paths to one resultA submitted paper of 40 questions enters three scoring paths at once. Objective questions are scored at submit against the answer key frozen at publish, with author-configured negative marking and proportional partial credit. Free-text and image answers are evaluated after submission against the author's reference answer, in APPLY mode which applies the score or PROPOSE mode which waits for a grader to confirm. Remaining subjective answers are graded by a person, with the reference answer beside the script. All three converge on one result record carrying the score, the rationale, which path produced it, and an audit entry for every change.Submitted paperDBMS · SEM 3 · 40 QUESTIONS28objective8free text4imageObjectiveAT SUBMITscored against the key frozen at publishnegative marking · proportional partial creditAI evaluatedAFTER SUBMISSIONcompared with the author’s reference answerAPPLY applies it · PROPOSE waits for a graderHuman gradedIN THE QUEUEthe reference answer sits beside the scripttwo graders mark the same wayOne result recordscoreper question and totalrationalewhy, in wordspathwhich one produced itauditevery change, with whoNOTHING ON THIS DIAGRAM RUNS DURING THE EXAM — EVERY PATH BEGINS AFTER SUBMIT
01

Automatic, at submit

Objective questions score instantly against the answer key frozen at publish — including author-configured negative marking and proportional partial marking on multiple-choice questions.

Three paths, one resultObjective questions score instantly against the frozen key. Free-text answers are evaluated by AI or graded by a person. All three converge on the same result record.ObjectiveINSTANT, AT SUBMITAI evaluatedAPPLY OR PROPOSEHuman gradedAGAINST A REFERENCEOne resultWITH A RATIONALEFULLY AUDITEDNEGATIVE AND PARTIAL MARKING APPLY
02

Manual grading

Subjective answers land in a grading queue with the author’s reference answer alongside, so two graders mark the same way.

Grading against a referenceSubjective answers land in a grading queue with the author’s reference answer alongside, so two graders mark the same way.The answerWHAT THEY WROTEAS SUBMITTEDThe referenceTHE AUTHOR’S INTENTSET WHEN WRITTENTWO GRADERS, ONE STANDARD
03

AI answer evaluation

Free-text and image answers can be evaluated against that reference answer, producing a score with a rationale. Per assessment, it runs in APPLY mode or PROPOSE mode, where a grader confirms before anything counts. Re-evaluation is audited, and monthly spend is capped per account.

Apply or proposeAI evaluation of a free-text answer either applies its score directly, or proposes it for a human grader to confirm. The mode is set per assessment.FREE-TEXTANSWERAUTO · APPLIEDPROPOSEGRADER CONFIRMSSET PER ASSESSMENTVS A REFERENCE ANSWER
04

After the exam, never during

AI evaluation runs asynchronously, after submission. The exam runtime does not call it — a provider outage or a rate limit cannot reach a live exam.

Evaluation runs after submissionAI evaluation runs asynchronously, after the candidate submits. The exam runtime never calls it, so a provider outage cannot reach a live exam.Candidate submitsTHE EXAM IS OVERQueue picks it upASYNCHRONOUSScore and rationaleAPPLIED OR PROPOSEDAN OUTAGE CANNOT REACH A LIVE EXAM

Bring one paper. We'll run it live.

Thirty minutes, your own exam, no slide deck.