Your skin knows more
than any wearable
ever will.
Tapas.one HealthOS is the world's first complete biosignal stack built on skin-native sensing — from a nano-cream applied like moisturiser to a Personal Health Twin that learns your unique biology. No rigid hardware. No needles. Just continuous, invisible health intelligence.
Layer 1: Nano-Cream Sensor
Not a watch. Not a ring. Not a patch. The Tapas nano-cream is applied directly to the skin surface — creating a biological interface that captures signals no external hardware can reach.
Invented by Pau Sabater, the formulation embeds biocompatible sensing elements that monitor 6+ physiological channels simultaneously, 24/7, without restricting movement or requiring charging. It reapplies daily like skincare.
Why skin-native sensing wins
Telemetry architecture
Layer 2: Signal Reader & Relay
A minimal hardware module — a small adhesive patch or smartphone-proximity relay — that reads the nano-cream's biosignals and transmits them wirelessly to the AI Health Engine.
Ultra-low power, no charging cables, no bulky hardware. It connects seamlessly to the Tapas.one mobile app and operates invisibly in the background — the only hardware in the stack, and the smallest possible one.
Layer 3: AI Health Engine
The intelligence core of Tapas.one HealthOS. A multimodal AI system that transforms raw physiological signals into actionable health intelligence — predictive alerts, anomaly detection, and real-time insights updated every 30 seconds.
Predictive Alerts
Detect stress spikes, inflammation onset, and recovery deficits before symptoms appear
Longitudinal Baselines
Build personal biological reference ranges that evolve with your body over time
Anomaly Detection
Flag deviations from your personal baseline with contextual explanations
Real-time Insights
Live HRV, stress index, and recovery score updated every 30 seconds
Cognitive Load Tracking
Correlate EDA and HRV patterns with mental performance and burnout risk
Clinical-grade Accuracy
Validated against gold-standard medical devices in controlled studies
Layer 4: Personal Health Twin
Your Personal Health Twin is a continuously evolving AI model of your unique biology. Unlike population-average wellness apps, it learns your individual patterns, rhythms, and responses — becoming more accurate and personalised with every day of data.
Over time, it predicts how your body will respond to stress, sleep deprivation, exercise, nutrition changes, and environmental factors — before you feel the effects. This is the long-term moat: a personalised health model that gets harder to replicate the longer a user stays.
The privacy promise
Your Health Twin data is yours. Stored with end-to-end encryption, never sold to third parties, and fully exportable or deletable at any time.
Real platforms. Real patients.
Powered by Tapas.one.
See how leading health AI platforms integrate Tapas.one's biosignal stack to deliver measurable clinical outcomes across neurorehabilitation, sleep medicine, and enterprise AI orchestration.
JustShowUp
·Precision Rehabilitation TechnologyJustShowUp's AI rehabilitation platform streams continuous EMG, computer-vision, and sentiment signals into Tapas.one's AI Health Engine. The Personal Health Twin builds a longitudinal recovery model for each stroke patient — adapting therapy difficulty in real time, flagging fatigue before it causes setbacks, and giving clinicians a single biosignal dashboard across all remote patients.
SleepFM
·Sleep Intelligence PlatformSleepFM's foundation model — published in Nature Medicine (2025) — decodes a single night's polysomnography to predict 130+ future diseases. Tapas.one's AI Health Engine routes each patient's nightly risk scores into their Personal Health Twin, creating a longitudinal disease-trajectory model that surfaces cardiovascular, neurological, and metabolic risks months before symptoms appear.
SleepFM.Life
·AI Sleep Inference PlatformSleepFM.Life is the open research inference layer of the SleepFM foundation model, published in Nature Medicine (2025). Its Leave-One-Out Contrastive Learning (LOO-CL) architecture encodes raw PSG signals — EEG, EKG, Airflow, SpO₂, EMG — into rich sleep representations for staging and disease prediction. Tapas.one ingests SleepFM.Life's per-patient risk scores for 13 time-to-event outcomes (dementia, AFib, heart failure, T2D, and more) into the Personal Health Twin, building a continuous disease-trajectory model that updates with every new night of sleep.
Apogeee.One
·Unified Intelligence PlatformApogeee.One — built by Eangelica Aton — orchestrates 12 NVIDIA NIM microservices across defense, clinical, telecom, and pharma verticals. Tapas.one serves as the intelligent routing and Semantic Cache layer: directing queries to the right NIM model, eliminating redundant GPU inference calls by up to 80%, and ensuring every vertical's AI output feeds back into a unified health intelligence stream.
SignalSys.Click
·AI-Powered Signal Processing PlatformSignalSys.Click is the core R&D infrastructure platform for Tapas.one's hardware and biosignal validation pipeline. Built by a global team of women engineers across Italy, the US, Hungary, El Salvador, and Tunisia, SignalSys provides FFT spectrum analysis, ECG/EMG/EEG MATLAB code generation, circuit simulation for biosensor hardware (op-amps, BJT/FET amplifiers), and real-time signal processing — all directly applicable to characterising nano-cream biosensor output and validating the Layer 2 reader/relay RF telemetry.
MedConnect.Team
·Elective Care & Medical Tourism NetworkMedConnect.Team connects patients with world-class elective care clinics across Türkiye — including Aesthetics & Dermatology, Bariatric Surgery, Dental, and Ophthalmology. Tapas.one's nano-cream biosensor integrates into every MedConnect procedure pathway: applied topically before surgery to establish a skin biomarker baseline, then worn continuously during recovery to stream glucose, cortisol, hydration, melanin, and cardiac signals to the clinic dashboard — even after patients return home across borders.
Validated in Digital Health by GoApercu
GoApercu is the operational proof-of-concept for AI-native healthcare intelligence. Its clinical accuracy benchmarks, HIPAA-compliant API infrastructure, and live patient monitoring capabilities establish the digital health baseline that Tapas.one's HealthOS is built to extend — from episodic clinical encounters to continuous, nano-scale biosensing.
How Tapas.one extends GoApercu: GoApercu operates at the clinical encounter layer — structured EHR data, episodic diagnostics, and model-driven treatment recommendations. Tapas.one adds the continuous biosensing layer beneath it: nano-cream sensors streaming real-time biomarkers into the Personal Health Twin, which feeds enriched longitudinal signals back into GoApercu's AI models for higher-confidence predictions.
99.7% Clinical Accuracy
GoApercu's AI diagnostics achieve 99.7% clinical accuracy across 15 live models (GPT-4 Medical, Claude Health, Med-PaLM 2), providing the ground-truth benchmark against which Tapas.one calibrates its Health Twin predictions.
<50ms API Response
45M+ API requests/month at sub-50ms latency via edge computing and intelligent caching. Tapas.one's Smart Router integrates with GoApercu's RESTful + GraphQL FHIR R4 endpoints to pull real-time clinical signals without latency penalty.
HIPAA · SOC 2 · ISO 27001
End-to-end AES-256 encryption, TLS 1.3, zero-knowledge processing, and full BAA coverage. GoApercu's compliance posture sets the security baseline that Tapas.one inherits for all PHI flowing through the HealthOS pipeline.
AI-Enhanced Clinical Dashboard
GoApercu's live dashboard surfaces patient risk scores (cardiovascular, respiratory, neurological), treatment recommendations with evidence levels, and AI accuracy trends. Tapas.one's Personal Health Twin ingests these signals to build longitudinal disease trajectories.
EHR Integration & Model Orchestration
Dynamic model orchestration with intelligent load balancing, ensemble learning, and specialty routing across 15 AI models. Tapas.one leverages GoApercu's EHR integration layer to unify structured clinical data with continuous biosensor streams from the Nano-Cream sensor.
24/7 Health Monitoring
Continuous patient vitals monitoring with real-time AI alerts. GoApercu's monitoring infrastructure validates Tapas.one's always-on sensing model — demonstrating that continuous, non-invasive health intelligence is both clinically viable and operationally scalable.
Signal Processing R&D — Powered by SignalSys
Tapas.one's hardware and biosignal R&D is validated on SignalSys.Click — an AI-powered signal processing platform built by a global team of women engineers across Italy, the US, Hungary, El Salvador, and Tunisia. SignalSys provides the analytical infrastructure for characterising nano-cream biosensor output, designing reader/relay circuits, and validating real-time biomedical signal pipelines.
Biosignal Processing Pipeline
Nano-Cream Biosensor
Skin-applied sensing medium captures ECG, EDA, sweat biomarkers, and microcirculation signals
Signal Acquisition
SignalSys processes raw biosensor output — FFT analysis, noise reduction, QRS detection, RMS envelope
AI Feature Extraction
Adaptive learning algorithms extract HRV, stress index, inflammatory markers, and sleep stage features
Tapas Health Engine
Processed signals feed the AI Health Engine and Personal Health Twin for longitudinal health intelligence
SignalSys Capabilities Used in R&D
Noise reduction, pattern detection, and anomaly identification with adaptive learning algorithms
RF, baseband, audio, biomedical, and industrial sensor signals — including nano-cream biosensor output
Interactive diode, BJT/FET amplifier, and op-amp simulations with transfer characteristics for biosensor hardware design
Bode diagrams, FFT visualization, and bandwidth analysis — used to characterise nano-cream sensor frequency response
Near real-time (<1s) with soft real-time (<100ms) for critical biosignal applications requiring continuous monitoring
Enterprise-grade RF signal analysis for the Tapas Layer 2 reader/relay module and BLE 5.3 transmission characterisation
Biomedical Signal Analysis — MATLAB Examples
[peaks, locs] = findpeaks(ecg_signal); heart_rate = 60 * length(peaks) / duration;
rms_envelope = movmean(emg_signal.^2, window); rms_envelope = sqrt(rms_envelope);
[b,a] = butter(4, [8 12]/(Fs/2)); alpha_band = filtfilt(b, a, eeg_signal);
40+ MATLAB examples covering ECG QRS detection, EMG muscle activity, and EEG brain wave processing are available on SignalSys.Click — all directly applicable to Tapas nano-cream biosensor signal characterisation.
Live Signal Analysis Demo
Try the SignalSys interactive demo directly — no login required. Run FFT analysis, adjust frequency parameters, and explore biomedical signal processing in real time.
Loading SignalSys demo…
Powered by SignalSys.Click — Next Gen AI-Powered Signal Processing · Built by Next Gen Engineers Worldwide
We use cookies
We use essential cookies to keep you signed in and optional analytics cookies to understand how Tapas is used — no advertising, no third-party tracking.
Essential
Always on — required for login & security
Analytics
Optional — anonymous usage stats