Tiger's Eye Detector is an entertainment-first lie detection app that analyzes a subject's recorded answer to a single question and produces an instant truth-versus-deception reading. Rather than relying on one signal, the app combines several categories of deception cues — vocal stress, eye movement, and micro-expression analysis — to generate a balanced, percentage-based verdict. While no app can truly read minds, understanding the cues the engine references helps you interpret each reading and have more fun interrogating your friends.

Vocal stress analysis listens for changes in pitch, pacing, hesitation, and word choice. When someone stretches the truth, their voice often tightens, speeds up, or fills gaps with filler words. The app transcribes the spoken answer and weighs specific phrases against known stress markers to estimate a vocal stress level between 0 and 100.

Eye movement and gaze tracking examines where the subject looks, how often they blink, and how stable their gaze remains. Inconsistent eye contact, rapid blinking, or gaze aversion are classic — though not definitive — indicators of discomfort or fabrication. A snapshot of the subject's face is captured at the end of each answer so the engine can reference visual behavior alongside the transcript.

Micro-expression scoring looks for fleeting facial movements that betray an emotion the subject is trying to hide. These involuntary expressions often last less than half a second and can reveal a mismatch between what someone says and what they actually feel. Combined with the vocal and eye signals, this produces the micro-expression score shown on every reading.

Each question is analyzed independently, so you can build a session one interrogation at a time and watch the truth and lie percentages shift as the conversation evolves. Results are for entertainment purposes only and are not scientifically validated — but they make for a dramatic, shareable party game. Below are the deception-detection methods the engine draws on, each with a short description and an estimated reliability reference.