InVivo
population-healthspan intelligencemeasurable value before and after devices

InVivo turns daily health signals into measurable population-health value.

InVivo is now built to identify preventable healthspan transitions, assign governed low-cost pilots, adapt the UX around a chosen modality focus, measure adherence and outcomes, and report aggregate value without crossing into diagnosis, prescribing, or medication changes.

Population Healthspan Value

Source-grounded events, governed pilot loops, consented cohort learning, and aggregate economic endpoints.

implemented suite
Model families
5

metabolic, cardio, recovery, inflammation, function

Pilot protocols
11

eligibility, adherence, outcomes, safety boundaries

Native parity
2

iOS Twin readiness and Android domain implementation

Cohort reporting
>=3

small-cell suppression before aggregate display

Federated learning
3 types

glucose, care-plan adherence, biosignal — secure aggregation only

Body systemsModality focusAutonomic/VagusTwin fidelityPopulation healthCarePlansTriager handoff11 pilot protocolsSource receiptsClinician packetMedicinesBiologicsSupplementsMealsSleepInsightBio ScanGlucoseKetonesLabsHRVGeneticsNutrigenomicsPrevention-firstHealthScoreCamera vitalsFederated AIOn-device Core MLAvatar twin
invivo / twin map / source receipts / 09:41
livebody systems / fidelity / clinician packet
Fidelity
64%
lab + phone
Systems
6of 9
covered
Receipts
18items
source-grounded
Ask twin
answer with receipts
Body map
9 subsystems
Packet
macOS export
Body-system map / current fidelity
Metabolic high · autonomic medium · liver/kidney missing
source receipts
Search answer
Why was recovery lower?
Answer cites sleep, HRV, meal timing, and Bio Scan confidence.
Packet receipts
ApoB 82 mg/dL
May 21
lab
Bio Scan HRV 44 ms
Today
phone
Mild bloating
Today
symptom
Nine-system map
Metabolic, autonomic, sleep, cardiovascular, immune, liver/kidney, gut, endocrine, and body-composition context stay organized.
Modality focus
Autonomic/Vagus, metabolic, sleep/recovery, labs/longevity, and CarePlan/recovery modes reorder the app without hiding other signals.
Fidelity, not hype
Phone-only, device-enriched, lab-verified, clinician-contextualized, and imaging-enriched tiers keep certainty honest.
Population health
Consented cohorts can measure preventable risk transitions, low-cost interventions, adherence, burden, and cost/utilization endpoints.
Search with receipts
Ask why something changed and see the sleep, meal, lab, device, or Bio Scan evidence behind the answer.
01 / Phone first

Useful even with no smartwatch, band, ring, or CGM.

InVivo starts with the device you already have. The phone can capture meals, sleep sessions, acoustic sleep context, camera-based wellness summaries, manual biomarkers, documents, and trends.

Meal journal from the phone

Take a photo, speak what you ate, or enter it manually. Adaptive meal logging keeps the flow clear in light or dark mode before any sensor is connected.

Phone-only sleep sessions

Use the iPhone overnight for a visible, opt-in SleepInsight session with charging readiness, microphone permission, night mode, source fidelity, correction labels, and derived summaries by default.

Bio Scan from the camera

Run a short phone-camera wellness scan for derived heart rate, respiration, HRV, signal quality, and confidence metadata. Raw video is not retained.

Manual glucose, ketones, and GKI

Enter fingerstick values or ketone readings by hand and InVivo still computes paired GKI, trends, and timeline context.

Lab reports and PDFs

Import blood tests, lipid panels, FHIR JSON, PDFs, and CSV files so clinical markers sit next to daily behavior.

Trends without hardware

Meals, symptoms, supplements, labs, phone sleep, and Bio Scan summaries are enough to start seeing patterns.

02 / Population health

InVivo can turn personal health tracking into measurable population-healthspan value.

The population-health opportunity is not another dashboard and not a single biological-age score. It is a consented loop: source-grounded events, preventable transition models, governed low-cost interventions, CarePlan packets, Triager handoffs, measured outcomes, economic endpoints, and honest safety boundaries.

Population Healthspan Value
Measurewhat actually improves

InVivo can support the hard part of population health: identifying preventable deterioration early, matching people to low-cost actions, measuring whether those actions worked, using SleepInsight to measure recovery regularity and breathing-load trends, following standardized CarePlans through consent-scoped packets, Triager handoffs, compliance and evidence, and reporting value without pretending sparse data is certainty.

Boundary
wellness + clinician discussion
Evidence
causal ladder
Pilots
11 protocols
Privacy
consent scoped
01
HealthEvent

source, time, confidence, consent

02
SleepInsight

fidelity, caps, corrections

03
Feature store

recency, missingness, device tier

04
Model card

purpose, inputs, boundaries

05
Pilot suite

11 protocols, safety, outcomes

06
CarePlan

packet, handoff, compliance

07
Cohort report

aggregate-safe healthspan value

Stronger claims require stronger evidence. InVivo can begin with personal n-of-1 experiments and prospective pilots, then move toward cohort validation only when calibration, missingness, fairness, drift, correction rate, and external validation are satisfied.

Source-grounded health events

Every signal can carry time, source, method, confidence, privacy class, consent scope, and source receipt metadata so population models are built from auditable evidence instead of unstructured wellness notes.

Preventable transition models

The value is not a vague health score. InVivo models transitions that can be prevented or delayed: worsening metabolic recovery, cardiovascular marker momentum, SleepInsight recovery debt, persistent inflammation context, and functional decline.

Governed intervention loops

The native model layer implements measurable protocols with eligibility, required inputs, preferred inputs, burden, cost, measurement windows, prohibited-use withholding, and aggregate-safe reporting.

CarePlan follow-up evidence

Standards-backed CarePlans now become consent-scoped Clinician Packets: included, withheld, stale and missing evidence; source receipts; Triager handoff context; observed effects; and reviewable signals for how a plan may need clinician review.

SleepInsight recovery evidence

Sleep becomes a source-grounded nightly recovery event: source fidelity, sanity validation, user corrections, breathing-load boundaries, HRV/resting-heart-rate context, and derived-feature cohort eligibility.

Cohort learning with consent

Individuals get personal feedback while consented cohorts can reveal which low-cost actions work, where data is missing, and which groups need better measurement before stronger claims are made.

Cost and utilization endpoints

Population health value is measured through avoided duplicated labs, completed follow-up, lower user burden, clinician minutes saved, low-cost intervention adherence, and healthier daily function.

Privacy-preserving boundaries

Population reporting is aggregate-safe, small-cell suppressed, consent-scoped, and explicit about what remains personal, what can be shared, and what is never diagnosis or medication advice.

Five model families / 11 implemented protocols

Model preventable transitions, not abstract wellness.

healthspan
Metabolic deterioration

Transition: Stable metabolic pattern -> poor glucose recovery or higher-risk lab band

Action
post-meal walking, earlier dinner, DPP-style prevention packet
Endpoint
glucose recovery, lab anchors, adherence, cost per improved risk state
Cardiovascular momentum

Transition: Stable marker pattern -> unmanaged BP, lipid, glucose, or recovery momentum

Action
home BP plan, lipid packet, follow-up preparation
Endpoint
duplicate tests avoided, clinician minutes saved, marker drift prevented
Recovery capacity

Transition: Stable recovery -> SleepInsight recovery debt, lower adherence, or fewer functional days

Action
sleep regularity, strain budget, meal cutoff, wind-down plan
Endpoint
source-fidelity sleep timing, HRV/resting HR, correction rate, healthy functional days
Inflammation context

Transition: Weak context -> persistent abnormal marker pattern needing clinician discussion

Action
repeat-lab packet, symptom timing, recovery-load review
Endpoint
unnecessary worry/testing avoided while preserving escalation
Function and frailty prevention

Transition: Stable function -> mobility decline, lower strength reserve, or falls context

Action
strength habit, protein adequacy, walking consistency, mobility check
Endpoint
functional days preserved and avoidable decline reduced
Value equation

Population Healthspan Value equals avoided risk transitions, improved functional capacity, better validated biomarkers, and healthier daily recovery minus intervention cost, unnecessary utilization, user burden, clinical burden, and governance risk.

Model families
5

metabolic, cardio, recovery, inflammation, function

Pilot protocols
11

every declared intervention has a measurable protocol

CarePlan loop
closed

packets, handoff, compliance, outcome review

SleepInsight FL
7 tasks

sleep/wake, acoustic, recovery, uplift, metabolic response

CarePlan FL
9-param

care-plan adherence round, consent-gated, secure aggregation

Native parity
2

iOS Twin readiness plus Android domain parity

Report rule
>=3

minimum cohort cell before aggregate display

Measurable pilot suite

Eleven protocols now make the portfolio assignable and reportable.

InVivo now has a measurable pilot layer for every declared population-healthspan intervention. Post-meal walking keeps the deepest glucose-specific outcome scoring; the full suite adds structured eligibility, CarePlan activity assignment, consent-scoped Clinician Packets, stale or withheld evidence, Triager handoffs, prohibited-use withholding, follow-up assessment, and small-cell-suppressed cohort reports across all five model families.

01
Eligibility

Check source coverage, protocol-specific required inputs, missing preferred inputs, readiness, and prohibited medical-use triggers before assigning a pilot.

02
CarePlan assignment

Select a declared protocol or standardized CarePlan activity with an action label, measurement window, required source context, and safety boundary.

03
Compliance evidence

Capture completion, burden, source count, SleepInsight correction labels, Clinician Packet readiness, stale or withheld evidence, and safety handoffs rather than assuming a recommendation was followed.

04
Outcome window

Measure response using predefined endpoints such as glucose recovery, BP coverage, SleepInsight sleep regularity, packet follow-up, functional days, duplicate testing avoided, or CarePlan goal movement.

05
Aggregate learning

Report only cohort-safe counts, adherence, missingness, response distribution, burden, economic endpoints, and plan-modification hypotheses with small-cell suppression.

Implemented protocol catalog

The first suite covers metabolic, cardio, recovery, inflammation, and function.

iOS Twin + Android domain parity
Metabolic14 days
Post-meal walk

meal adherence and 30/60/120-minute glucose response

Metabolic7 days
Earlier dinner

overnight recovery, morning glucose or ketone context

Metabolic90 days
DPP-style prevention

packet review, activity minutes, weight/waist or lab follow-up

Cardio14 days
Home BP tracking

reading consistency, source provenance, follow-up readiness

Cardio30 days
Lipid clinician packet

clinician review, duplicated tests avoided, source-document utility

Recovery7 days
Sleep regularity

source-fidelity sleep timing, HRV/resting HR, correction rate, recovery debt, functional-day reports

Recovery7 days
Strain budget

accepted suggestions, correction rate, next-day recovery

Inflammation45 days
Repeat-lab context

repeat-lab completion, escalation preserved, duplicated explanation avoided

Inflammation14 days
Recovery-load review

symptom resolution, recovery movement, lab-anchor timing

Function28 days
Strength habit

session adherence, walking consistency, soreness, functional-day report

Function30 days
Mobility check

mobility-check completion, walking trend, escalation completion

Who benefits

Health systems

Prepare better pre-visit context, reduce duplicated history taking, and identify which standardized CarePlans produce measurable follow-through using Clinician Packets before escalation.

Research partners

Run consented, source-grounded observational pilots with explicit missingness, device tier, standards evidence, subgroup calibration, and outcome-window capture.

Employers and payers

Evaluate aggregate healthspan programs through CarePlan packet readiness, compliance, burden, avoided duplicate testing, functional days, and privacy-safe cohort summaries rather than surveillance.

Individuals and clinicians

Keep personal insight useful while packaging source receipts, medicine/supplement context, and clinician questions when the next step belongs in care.

Governance and boundaries
No diagnosis, treatment, insulin dosing, medication adjustment, or disease-prevention guarantee.
Every model family reports source coverage, missing inputs, confidence, next validation step, and prohibited uses.
Medicine and supplement context is database-backed when possible, with manual fallback clearly labeled.
SleepInsight summaries carry source fidelity, hard sanity caps, correction labels, and non-diagnostic breathing-load boundaries before personal or cohort use.
CarePlan status is evidence-backed: InVivo can prepare patient-owned packets and Triager handoff context, while clinician verification, medication, order, photo, high-risk tasks, documentation, export, and plan mutation require the care workflow.
Federated learning is opt-in per model: care-plan adherence, glucose-state, and biosignal updates are combined only through secure aggregation, so no individual contribution is ever seen in the clear and raw signals never leave the phone.
Causal evidence moves through plausibility, retrospective checks, target-trial emulation, n-of-1 pilots, prospective cohorts, pragmatic trials, and external validation.
Outputs can be withheld when data is missing, unsafe, too sparse, or outside the wellness/clinician-discussion boundary.
03 / Fidelity + receipts

Twin Fidelity tells you how much the model actually knows.

Instead of pretending every signal has the same certainty, InVivo labels which evidence tier is active and keeps the original source visible.

Twin Fidelity
64%lab + phone + partial device context
not a health score
Phone-onlyactive from day one

Meals, Bio Scan, SleepInsight phone sessions, symptoms, manual readings, and journal context.

Device-enrichedoptional precision

Apple Health, CGMs, watches, rings, chest straps, scales, and sync recency.

Lab-verifiedsource grounded

Blood tests, lipids, CMP, inflammation markers, PDFs, FHIR JSON, and CSV imports.

Clinician-contextualizedcare conversation

Symptoms, medicines, supplements, protocols, visit notes, and questions for review.

Imaging-enrichedfuture fidelity

Radiology reports, DEXA, ultrasound, MRI, and third-party deep-assessment imports.

Body search / source-grounded answer
“Why was my recovery lower this week?”

Likely contributors: shorter sleep, lower Bio Scan HRV, later meal timing, and a recent bloating symptom. This is a pattern summary from your timeline, not a diagnosis.

Sleep
7h 42m
phone session
HRV
44 ms
Bio Scan
ApoB
82 mg/dL
lab PDF
GKI
3.0
manual paired reading
Data honesty layer

Every explanation can point to source type, modality, date, confidence, and whether the signal came from a phone, device, lab, clinician note, or imported report.

04 / Try it now

Estimate your heart risk, right in your browser.

Real AHA PREVENT science (Khan et al., 2023), computed entirely on this page — move a slider and the number updates, and nothing is ever sent anywhere. The same on-device engine the app ships.

Age50
Blood pressure (top number)124 mmHg
Total cholesterol200 mg/dL
HDL (“good”) cholesterol50 mg/dL
Body-mass index26
10-year total cardiovascular risk
2.2%Low
30-year: 15.2% · 10-yr ASCVD: 1.4%
Computed in your browser. Nothing is sent anywhere.

Educational cardiovascular-risk awareness (AHA PREVENT, Khan et al. 2023). Not a diagnosis — discuss prevention with a clinician.

Track this continuously with InVivo
05 / Whole body

One app. Your whole body.

Dozens of health models work together on your device, across every major system — so your insight is broad, personal, and private by design.

45+
health models
13
body systems
100%
on your device

Heart & circulation

Cardiovascular risk, ECG insight, and early-warning triage.

  • 10 & 30-year heart-risk
  • ECG intervals & morphology
  • Acute-concern triage
  • Rhythm screening flags
  • Heart-rate variability
On-devicePrivateFederatedScreening

Metabolic & glucose

Glucose patterns, forecasting, and metabolic balance.

  • Glucose variability
  • Short-horizon forecasting
  • Metabolic balance
  • Fasting stages
  • Keto & GKI context
On-devicePrivateFederated

Sleep & breathing

Sleep quality, snoring, and breathing-load screening.

  • Sleep insight
  • Snore analysis
  • Apnea screening
  • Sleep regularity
On-devicePrivateScreening

Autonomic & recovery

Nervous-system balance and recovery capacity.

  • Vagal / HRV tone
  • Orthostatic response
  • Recovery & strain
  • Stress load
On-devicePrivate

Aging & vitality

Wellness-grade biological- and functional-age context.

  • Metabolic age
  • Functional age battery
  • Whole-body HealthScore
  • Longevity pace
On-devicePrivate

Genomics

Educational genetic context from your own file.

  • Polygenic risk context
  • Pharmacogenomics
  • Nutrigenomics
On-devicePrivate

Gut health

Digestive patterns and microbiome report context.

  • Bristol stool insight
  • Microbiome report context
On-devicePrivate

Mind & mood

Validated mental-wellness screeners and resets.

  • PHQ-9 / GAD-7 / WHO-5
  • Symptom triage
  • Somatic reset
On-devicePrivateScreening

Camera & vitals

Hands-free wellness signals from your phone camera.

  • Pulse (rPPG) trend
  • Respiration context
  • Tongue & eye wellness
On-devicePrivate

Nutrition

Effortless meal capture and nutrition estimates.

  • Meal photo & voice
  • Ingredient detection
  • Nutrition estimates
On-devicePrivate

Environment

How your surroundings shape your health.

  • Place context
  • Air, pollen & daylight
  • Digital & stress load
On-devicePrivate

Clinical & labs

Make sense of labs and prepare for your clinician.

  • Lab & document import
  • Clinician packets
  • Care-plan adherence
On-devicePrivateFederated

Your digital twin

A source-grounded twin that cites or stays silent.

  • Evidence-grounded answers
  • Personal voice & avatar
  • Proactive nudges
On-devicePrivate

Wellness, education, and clinician-discussion tools — not a diagnosis. Stronger outputs stay gated until they pass independent validation.

06 / Your data

Your data never leaves your device.

Most health apps send your data to their servers. InVivo flips it: the models come to your data, not the other way around.

Runs on your phone

Every model analyzes your raw signals locally. Your ECG, glucose, photos, labs, and genome never leave the device.

Learns without looking

If you opt in, your phone shares only tiny, math-masked model updates — never your data. They’re encrypted and combined with thousands of others, so your contribution can’t be singled out. Even we can’t see it.

Everyone’s model improves

The smarter model comes back to your device. Your data stayed with you the whole time.

Never leaves your device
  • Raw ECG & heart rhythm
  • Glucose & metabolic data
  • Photos & meal images
  • Lab results & documents
  • Your genome
  • Voice & face
The only thing that can ever be shared
  • Anonymous model updates — math, not data
  • Encrypted, with statistical noise added
  • Blended with thousands, so you can’t be singled out
  • Only if you opt in — revocable anytime

Nothing is shared. Everything stays on your device.

On-device by defaultEncrypted when sharedDifferential privacyConsent-gatedDelete everything, anytime
07 / Heart dynamics

A healthy heart is not a metronome.

Plot each heartbeat interval against the next and a healthy heart traces a broad “comet” — and that breadth is complexity. Much of cardiac disease is a loss of complexity: the dynamics becoming too regular, or dissolving into turbulence. It's the nonlinear-dynamics story behind InVivo's chaos and Poincaré biomarkers.

Physiologic
Pathologic
Life-threatening

Normal sinus rhythm

Phase space · stable limit cycle

A stable limit cycle — the orbit the heart snaps back to after a perturbation.

Healthy HR variability

Return map · fat comet (high complexity)

A broad, fractal “comet.” A healthy heart is not a metronome — complexity is health.

Atrial fibrillation

Return map · diffuse cloud (irregularly irregular)

The comet collapses into a round cloud — almost no correlation between beats.

T-wave alternans

Phase space · period-2 (period-doubling onset)

A period-2 orbit (long-short-long-short) — the first whisper of chaos.

Ventricular tachycardia

Phase space · fast monomorphic loop

A single fast re-entrant circuit — still periodic, just dangerously fast.

Ventricular fibrillation

Phase space · strange attractor (deterministic chaos)

A strange attractor — bounded chaos that never repeats, exquisitely sensitive.

Illustrations of cardiac dynamics, not readings of a heart. The two return maps are stochastic surrogates and the fibrillation panel is a low-dimensional stand-in for high-dimensional tissue chaos — the shapes (comet vs. cloud, limit cycle vs. strange attractor) are the genuine signatures. InVivo plots your real Poincaré comet on-device.

Recommended device

For heart dynamics, a Polar H10 chest strap.

The chaos, Poincaré, and Cardiac Dynamical Deviation biomarkers read the beat-to-beat geometry of your heartbeat intervals — so they need continuous, electrode-grade R-R. Optical wrist sensors sample intermittently and drop beats under motion: excellent for trends, but not for the beat-to-beat structure these features depend on. A Polar H10 streams near-ECG interbeat timing continuously, and InVivo reserves its heart-dynamics readouts for that class of signal.

Any continuous chest-strap source works — we simply find the Polar H10 the most reliable. Wrist and finger sensors still power the rest of InVivo.

08 / Experiments

Most health hacks do not survive measurement.

InVivo Experiments turns wellness folklore into small, safe, falsifiable personal experiments — then aggregates the results through federated evidence. A truth engine, not a recommendation engine.

Folklore → falsifiable question
You say

“Magnesium improves my sleep.”

InVivo tests

Does magnesium glycinate nightly for 14–28 days improve your sleep latency by at least 15 minutes versus control — beyond your normal week-to-week variation?

Design
28-day crossover
MWE
15 min faster onset
Every claim passes five gates
01Safe enough to test?
02Outcome measurable?
03Falsifiable in time?
04Placebo reducible?
05Aggregates privately?
First five claims
Post-meal walk and glucoseCaffeine cutoff and sleepConsistent wake time and recoveryMagnesium and sleepBlue-light glasses and sleep
How the falsification engine works
preview / living twin / clinician packet

Start with your phone. Build fidelity as your evidence grows.

InVivo is built for people who want a serious personal health system without being forced to buy a stack of hardware first. Add devices, labs, imaging reports, and clinician context when you want higher fidelity and better source-grounded answers.

Preview intake is handled directly until a public signup endpoint is live.