/ products / anti-cheat sandbox
Coding Assessments with Evidence, Not Guesswork.
A full-featured Monaco code editor with layered integrity signals: server-timed attempts, keystroke and paste replay, cross-candidate similarity, tab-change logging, and a voice defense of the submitted code. Code runs in sandboxed containers. Signals flag risk for a human reviewer; nothing is auto-rejected.
Keystroke & Paste Replay
Replays the editing session and checks that the final code was actually typed in the editor. Large pastes are logged and shown to the reviewer.
Server-Timed Attempts & Tab Logging
Attempt timing is enforced server-side; focus changes, tab switches, and full-screen exits are logged with timestamps.
Cross-Candidate Similarity
Token-fingerprint comparison surfaces submissions that closely match other candidates, so reviewers can spot shared or leaked solutions.
Multi-Language Container Isolation
Candidate code runs on our self-hosted Piston sandbox in resource-limited, ephemeral containers across popular languages.
Automated Hidden Test Runner
Hidden tests probe edge cases, boundary conditions, and concurrency traps. Brownfield tasks ask candidates to fix or extend an existing, subtly buggy codebase.
Reviewer Evidence Pack
Code, test results, keystroke replay, similarity matches, and the voice-defense transcript in one place. In AI-allowed assessments, the candidate's AI prompts are logged too.