Collect the signal.Keep identity optional.
Devices turns supported health-device readings into validated, identity-light observations for care, research, and operational workflows—without username and password friction.
Designed for collection and workflow support. This public surface uses representative data and does not diagnose or display live patient records.
Pulse oximeter
OXY-S4 / 7F3A
Oxygen
97%
Pulse
72bpm
Perfusion
4.8%
Periodic SpO₂ and pulse readings retain measurement time, signal quality, and perfusion context.
AI-assisted ECG review
Connected-device ECG data, transformed into an AI-analyzed report.
A PulseBit device captures the ECG session and Platforms preserves the signal, provenance, and recording context. AI analyzes heart-rate patterns, rhythm events, variability metrics, and selected fragments into a structured report for qualified clinical review—not autonomous diagnosis. This de-identified sample exposes no patient identity.
AI report conclusion
Normal sinus rhythm
Reference-only algorithmic finding. Clinical interpretation remains with a qualified professional.
Minimum
56bpm
Average
66bpm
Maximum
73bpm
Recording
3m 11s
Heartbeats
152
Devices proof surface
Many devices.One governed view.
See what arrived, which device produced it, whether it is fresh, and what requires review—without asking operators to interpret raw payloads or switch between manufacturer utilities.
Supported reading → governed observation
A visual model of the faceless collection path.
Faceless collection path
For supported devices, the path from reading to normalized observation avoids staff transcription, per-device export handoffs, and spreadsheet reconciliation as the default way work gets done.
Operator experience
Clarity beforeclinical context.
The operational view explains collection state and next action. It does not silently turn device data into diagnosis.
Faceless collection
Supported readings arrive through a gateway without staff transcription, patient-facing login, or device-by-device exports.
One observation model
Different vendor payloads normalize into consistent pulse oximetry, blood pressure, ECG-session, and weight observations.
Visible data quality
Fresh, delayed, stale, unavailable, and under-review states remain explicit instead of disappearing into technical logs.
Governed handoff
Provenance, mapping, contract version, and masked device references travel with the observation into downstream systems.
Telemetry and trust
Technical detail,bounded by design.
Each observation preserves the source, mapping, measurement time, freshness, and review state needed for audit and integration. Patient identity stays outside the default record shape.
Source-to-object mapping
Customer-safe observation
{
"observationType": "pulse_oximetry",
"sourceDeviceFamily": "OxyS4",
"deviceRef": "masked-device-7f3a",
"measurementTime": "2026-07-20T09:14:00-07:00",
"freshnessState": "fresh",
"values": { "spo2Percent": 97, "pulseRate": 72 },
"identityState": "not_joined",
"reviewState": "ready"
}AI can
Summarize governed observations, freshness, and documented exceptions for non-diagnostic review.
AI cannot
Diagnose, choose treatment, claim emergency monitoring, or infer identity from an identity-light record.
Human control
Clinical interpretation, policy, exception resolution, and identity binding remain governed human decisions.
Evaluation
Start with abounded workflow.
A useful pilot begins with supported device families, representative readings, clear identity boundaries, and acceptance criteria for the workflow that needs the output.
Intake and review
Clinics
Reduce transcription and app switching while giving staff a clear view of current readings, gaps, and exceptions.
Repeatable collection
Research teams
Preserve source lineage, timestamps, mapping, and FHIR-compatible structure for study review and export.
Shared visibility
Medical groups
Give clinical and operations teams the same governed observation context without replacing clinical judgment.
Remote follow-up
Home health
Separate fresh, delayed, unavailable, and under-review readings across distributed patients and devices.
Qualified review artifacts