SwallowSafe AI shifts dysphagia support from static, preset drinking cups into patient-responsive, closed-loop fluid pacing by physically linking throat-motion feedback to adaptive fluid delivery.

Many rehabilitative swallowing-risk patients can drink orally, but ordinary fixed-volume aids do not respond when fatigue, delayed swallow recovery, or poor coordination appears during a meal. SwallowSafe AI is a clinician-prescribed assistive device that extends a speech-language pathologist's safe sip limits into daily drinking without claiming to diagnose aspiration or replace clinical judgment.

The system achieves this through two paired components:

• Biometric Kinematic Collar: A soft, non-adhesive wearable. A sealed, wipe-clean, inductive-charging sensor lozenge slides into a washable sleeve with an anatomical centering contour to reduce sensor rotation. It measures local throat motion and low-frequency vibration during active drinking. It is acoustically blind, so it cannot record intelligible speech or ambient conversation.

• Adaptive Cup Lid: Liquid meters through a drop-in, multi-cycle sterilizable liquid silicone rubber cartridge. A tamper-resistant, spring-loaded external flow-control actuator uses an auto-open loading latch for simple assembly, then enforces discrete bolus metering by releasing one programmed volume at a time. If power is lost, the actuator returns to fail-safe restricted flow. Liquid never touches electronics or internal micro-valves. An angled gravity-assist floor and elastomeric air-equalization vent support smooth delivery during safer chin-tuck drinking postures.

During clinical programming, an SLP sets safe limits such as bolus volume, consistency range, flow speed, and a custom tolerance band. The patient performs supervised swallows to create a baseline waveform profile stored locally on the collar. When the cup is raised, proximity pairing wakes the system. If the correct collar is absent, the cup defaults to conservative restricted flow.

During each drinking window, an edge-processed template-matching algorithm compares the current swallow-motion pattern with the baseline and tolerance band. It filters rhythmic chewing motion and uses sip-confirmation gating to exclude empty-cup dry swallows. The output is not a diagnosis; it is a physical pacing command. The cup releases the next bolus only when the programmed clearance interval and acceptable throat-motion pattern are met. If fatigue, delay, or weak motion appears, the actuator reduces or slows the next bolus. A subtle visual pacing indicator shows the patient the cup is intentionally pacing, preventing frustrated over-tilting or shaking. Repeated concerning patterns trigger a caregiver check, while minor variations are handled locally to avoid alarm fatigue. The system logs hydration volume and pacing-assistance frequency for caregiver workflow and therapy documentation.

SwallowSafe AI is designed from molded medical-grade plastic, washable collar sleeves, reusable silicone fluid-path cartridges, and a tamper-resistant detachable electronic drive pod. Its primary markets are rehabilitation hospitals, skilled nursing facilities, home-health programs, and outpatient swallowing therapy clinics. By reducing exposure to oversized or poorly timed sips associated with aspiration risk, SwallowSafe AI addresses a known contributor to costly pneumonia readmissions in dysphagia-risk patients while supporting safer independence, better tolerated hydration, and dignity at the dining table.