Manual paper builder
Build a paper section by section with MCQ, short and long answer questions, marks and instructions, then finalize it for scheduling.
AI CBT runs secure computer-based examinations for universities and colleges, from building and scheduling the paper to live proctoring, evaluation and published results. Every institution works in its own private workspace.
The paper, the sitting and the evidence live in one system, with a single well-defined handover between preparing an exam and running it.
From a blank paper to a scheduled sitting with its candidates, invigilators and rules in place.
From the first candidate signing in to a published result, with the integrity record to defend every mark.
Built for university and college examination bodies that need the paper, the exam and the evidence to live in one governed system.
Build a paper section by section with MCQ, short and long answer questions, marks and instructions, then finalize it for scheduling.
Build an exam from a course and semester, then schedule each subject sitting with its own date, window, duration and result policy.
Scheduling a paper copies it into an immutable snapshot. Editing the source paper afterwards cannot alter an exam that candidates have already sat, so results stay reproducible.
Answers save continuously against a server-held clock. A dropped connection resumes where it left off, and extra time can be granted to an individual candidate.
Tab switches, focus loss, missing or multiple faces and phones in frame are recorded against the attempt with severity and evidence.
Invigilators watch every candidate in a sitting, ranked by risk so the candidates who need attention come first, and record attendance for their duty.
Candidates scan a QR code and photograph their answer sheets page by page from a phone, attached straight to their attempt.
Objective answers mark themselves. Descriptive answers queue for an examiner, who can ask for an AI suggestion but always awards the mark.
Results are computed, stored and published deliberately, with per-question analysis to show which questions candidates found hardest.
Six stages, each owned by a role that already exists in your examination structure.
Write the paper in the manual builder: sections, questions, marks and instructions.
Admin · HOD · ExaminerSign the paper off so it becomes available to schedule.
Admin · HODAttach the paper to a sitting; it is frozen into a snapshot at that moment.
AdminTimed, proctored attempt served from the snapshot and saved continuously.
Candidate · InvigilatorObjective answers are marked automatically; descriptive ones by an examiner.
ExaminerCompute results and release them on your own schedule.
AdminAn examination system is judged on the worst day, not the average one. These are the guarantees that matter when a mark is disputed.
One university's papers, attempts and results are never reachable from another's workspace.
Once an attempt is submitted, what the candidate saw and answered can always be reconstructed exactly.
Administrative actions, including support impersonation, are logged with actor, address and outcome.
A proctoring flag only decides whose evidence is reviewed first, and an AI marking suggestion is only a suggestion. No candidate is failed automatically.
Your institution gets its own private workspace, its own address and its own sign-in policy. Sign in to get started, or talk to your platform administrator about onboarding.