In-Browser AI Proctoring
During the exam, AI models running in the candidate's own browser watch the camera and microphone for a missing face, extra people, a phone, looking away, head turning and talking.
Run timed online assessments with AI proctoring that works inside the candidate's browser. Camera and microphone checks flag suspicious moments, sessions are recorded, and your reviewers make the final call.
Secure Exam Proctoring is a multi-organisation online exam platform with AI proctoring built in. Organisations build a question bank, create and publish exams, and assign them to candidates or whole batches. Candidates sit a timed exam in their browser after a short system check of their camera, microphone and face position.
While the exam runs, AI models on the candidate's own computer watch for issues such as no face, several people, a phone, looking away or talking. Video is never streamed live: only flagged events and the recording are uploaded. Staff then review results, AI events and recordings in one place, and a person, not the software, decides whether an attempt is cleared.
Everything needed to author, run, monitor and review online exams.
During the exam, AI models running in the candidate's own browser watch the camera and microphone for a missing face, extra people, a phone, looking away, head turning and talking.
Choose which checks to run, how sensitive each one is, whether to record video and audio, and how many days to keep recordings.
Attempts are recorded and uploaded in short chunks, with evidence snapshots on flagged moments. Clicking an event in the timeline jumps the video to that moment.
Flags and risk levels are advice for reviewers. An evaluator records the decision (Cleared, Suspicious or Violation) with notes, and past decisions are kept as history.
Single choice, true/false, multiple select, short answer, long answer and coding questions, with marks, negative marks, difficulty, tags and images.
Candidates write C#, Python or JavaScript. Programs run against your test cases on a separate, isolated code runner, and hidden test cases stay secret.
Keep a reusable question bank. Import up to 1,000 questions from an Excel template, with a preview that highlights every wrong cell before anything is saved.
The timer comes from the server, answers save automatically, and a page refresh or short internet drop resumes the exam where the candidate left off.
See who is taking an exam right now, and their alerts as they happen.
Run separate organisations on one platform, with Super Admin, Admin, Instructor, Evaluator and Candidate roles. One organisation can never see another's data.
Every attempt with score, pass/fail, time taken, number of AI flags and risk level. Filter it or export it to CSV for Excel.
Every recording view, download and delete is logged, and recordings are removed automatically after the retention period (90 days by default).
Write questions in the editor or import them from Excel, with marks, negative marks and tags.
Set the duration, time window, attempts, pass mark, navigation rules and proctoring checks, then assign candidates.
After a system check and face calibration, the candidate answers while answers save and the AI watches.
Open each attempt to see answers, coding results, the AI event timeline and the recording, then record your decision.
Each organisation gets its own separate users, exams and results.
Conduct online internal and semester assessments with written answers marked by evaluators.
Run hiring assessments with coding questions scored automatically against test cases.
Assign exams to whole batches and review results, flags and recordings per candidate.