Hospital‑grade vital signs from $44 of WiFi hardware.
Continuous bedside monitoring costs $3,000 per room and gets reserved for the ICU. ViFi delivers wellness-grade contactless monitoring for 1/68th the cost, validated against a Polar H10 chest-strap at 4.15 bpm cross-session MAE.
What's been shown
The hardware works. It's been tested.
4.15 bpm cross-session HR error against a Polar H10 chest-strap reference, on a 2-chip array of $25 ESP32-S3 boards. PhaseBeat (the academic reference using $500 WiFi cards) reported 1.5 bpm. ViFi: 4.15 bpm on 20× cheaper hardware.
ViFi delivers 96% of PhaseBeat's accuracy at 4.4% of its cost. Against a clinical bedside monitor, it's 1/136th the per-room cost.
Why this exists
Most hospital patients are unmonitored more than 23 hours a day.
Outside the ICU, across 2.3 million U.S. hospital and skilled nursing facility beds, a nurse takes vital signs every 4 to 8 hours. The rest of the time, no one is watching. Sepsis spikes, post-surgical complications, opioid-induced breathing problems. Most of these happen in those gaps and only get caught at the next round.
Continuous monitoring exists, but it costs about $3,000 per bed and means strapping wires and electrodes to people who often pull them off. So hospitals reserve it for the ICU. The much larger floor of ordinary patients gets nothing.
From the founder
I'm building ViFi because it's the platform I wish my mother had when she spent a month in the hospital with sepsis. The wires, adhesives, blood-pressure cuffs, and 3 a.m. light checks meant to keep her safe were also what kept her from resting and healing.
How it works
WiFi signals bend around bodies. ViFi listens to how.
Two small chips in the room exchange WiFi signals back and forth. Every breath you take and every heartbeat slightly changes how those signals travel through the air. ViFi reads those changes as vital signs. No contact, no wearables, no line of sight required.
[ESP32-S3 TX] ─────── 192 subcarriers ───────► [ESP32-S3 RX]
antenna (chest perturbs path) antenna
│
│ USB serial @ 921600 baud
▼
csi_capture.py
│
▼
first_capture_report.py
│
▼
HR / RR predictions vs Polar H10 ground truth Signal pipeline
- Variance-rank top-K subcarriers
- Butterworth 0.1–3 Hz bandpass
- 4× zero-padded FFT
- Parabolic peak refinement (cardiac and respiratory bands)
- 9-dimensional feature vector → XGBoost regressor
Where the science comes from
The signal-processing approach comes from peer-reviewed academic work: PhaseBeat (INFOCOM 2017), FullBreathe (UbiComp 2018), ResBeat (2020). ViFi runs it on $15 commodity ESP32 chips instead of $500 research-grade WiFi cards, with the calibration, signal quality checks, and audit trail an FDA reviewer expects.
What's next
The path to a hospital-approved medical device.
Today ViFi is a pre-clearance research prototype, not a medical device. Three dated milestones stand between this and a contactless monitor running over a hospital bed:
-
Summer 2026
Multi-subject, multi-room HR validation
Validate heart rate across 10+ people in 3+ rooms against the same chest-strap reference. Target: under 3 bpm cross-subject MAE on subjects the system has never seen.
-
Spring 2027 (funding-dependent)
First hospital pilot
Wellness-grade install in 5–10 rooms alongside the existing standard of care. Patient data stays on-site and de-identified; no clinical decisions are made from ViFi during the pilot.
-
Late 2027 (funding-dependent)
FDA 510(k) clearance
Clinical validation study at an academic medical center, then file safety + accuracy data for FDA 510(k) Class II clearance. This is what makes ViFi legally usable for medical decisions in U.S. hospitals.
For hospitals
Pilot sites wanted.
ViFi is in conversation with medical/surgical floor leaders, biomedical engineering teams, and clinical research offices interested in hosting a wellness-grade pilot of contactless monitoring.
A pilot looks like this:
- 5–10 rooms instrumented with 2 ESP32-S3 boards each
- 30–90 day duration, IRB review where applicable
- Runs alongside your existing monitoring; no clinical decisions are made from ViFi data
- Fully de-identified, on-premise data handling
- Output: a measurement study comparing ViFi against your existing reference