The itch Levine's NEAT (Non-Exercise Activity Thermogenesis) research stopped us mid-scroll: habitual fidgeters can burn around 350 kcal more per day than sedentary people — not through any structured exercise, just unconscious leg motion. The frustrating part was having no way to know which category you fall into on any given day, or even which hour. So we built BinBot. How the detection works The core is a 100Hz CoreMotion capture pipeline run through a 3–9Hz IIR bandpass filter. Human fidget motion tends to live in that frequency window. Walking sits lower — around 0.5–2Hz — and ambient vibration from cars or trains is irregular enough to mostly fall outside the band. A five-layer rhythm classifier runs on top of that to further reduce false positives. Three motion modes are identified and tracked separately: vibrate, jump, and spin. All counts write to HealthKit.…