Whether you're grading essays, hiring talent, or collecting data, the problem is the same: AI spam. The solution is provenance.
Analyzing invisible metadata to validate human effort.
We measure the milliseconds between keys. Humans have a 'heartbeat' when they type—bursts of speed and pauses for thought.
We calculate a weighted score (0-100) based on paste frequency, tab switches, and editing behavior. Green means Human.
We detect unnatural text insertion blocks. Our engine differentiates between 'move text' operations and 'paste from ChatGPT' dumps.
We track window blur events and tab switches. Know if a candidate left the assessment page to Google the answer.
Watch the document creation in real-time. See every backspace, pause, and edit to verify the thought process.
We validate the hardware environment to detect headless browsers, bot farms, and automated testing tools.
We monitor shifts in vocabulary and tone. Sudden improvements in writing quality flag potential AI assistance.
We verify IP consistency and flag impossible travel scenarios or suspicious VPN usage during assessments.
Face tracking locally analyzes eye-out-of-focus movement to detect screen copying. We DO NOT record or transmit video. All analysis happens on-device.