Behavioral
Short-horizon exposures
· Active energy · Dietary energy · Carbohydrate · Sodium · Alcohol · Daylight · Mindfulness
Hours → days
Explainable longitudinal health intelligence.
Health intelligence
you can interrogate.
A scientific workspace for longitudinal self-experimentation. BioMIR turns selected Apple Health data into an explainable longitudinal view of what is changing, which modeled contributors are shaping the age-equivalent Δ-year output, how strongly the available data support interpretation, and whether the trajectory persists over time.
No dedicated BioMIR wearable required. Apple Health integration.
Core health calculations on device. Contributor-level interpretation.
Today · iPhone1 / 7
Swipe to explore · Tap to enlarge
Actual BioMIR screens · Demo data
BioMIR analytical architecture
One longitudinal record.
Two analytical views.
BioMIR organizes longitudinal data by analytical role and characteristic timescale. Adaptive BioAge integrates Behavioral and Functional contributors with a Cardiometabolic baseline, while Clinical Long View preserves episodic clinical-panel context separately.
View 1 Lifestyle Short View Daily longitudinal synthesis
View 2 Clinical Long View Episodic longitudinal synthesis
Functional
Responsive physiologic capacity
· HRV · Resting HR · VO₂max · Total sleep · Deep sleep
Days → months
Cardiometabolic
Slow-moving baseline state
· BMI · SBP · FBG
Weeks → months
Clinical panel
Episodic reference state
· Alb · ALP · BUN · Cr · CRP · HbA1c · SBP · TC · Glu · Lymph% · MCV · RDW · WBC
Episodic · dated panels
Adaptive BioAge
Integrated longitudinal model
Hierarchical composite with an age-calibrated cardiometabolic baseline and bounded, asymmetric Functional and Behavioral Δ-year contributions. The output is expressed in age-equivalent Δ-years relative to chronological age, with negative values representing age-equivalent reward and positive values age-equivalent penalty, while retaining contributor attribution over time.
Klemera–Doubal Method (KDM)
Multivariate biological-age estimator combining age-calibrated biomarker regressions with chronological age.
Levine PhenoAge
Mortality-trained multivariable biomarker score mapped through a Gompertz function to an age-equivalent phenotype.
Cadence depicts characteristic observation and interpretation timescale—not a causal sequence, fixed biological deadline, or prescribed testing interval. Cardiometabolic state is interpreted longitudinally; Clinical Long View is derived from episodic, dated clinical panels and remains computationally separate from Adaptive BioAge.
What sits beneath the model
Fifteen inputs.
Three distinct roles.
The architecture is intentionally compact: selected signals are grouped by how the model interprets them—not as 15 equal votes.
017 exposure inputs
Behavioral
Modifiable routines and exposures enter Adaptive BioAge as bounded contributors rather than direct measurements of physiological state.
025 physiology inputs
Functional
Autonomic regulation, restorative physiology, and cardiorespiratory capacity provide responsive physiological context across different measurement and adaptation timescales.
033 baseline inputs
Cardiometabolic
A slower-moving age-equivalent baseline anchors the daily composite, with original observation dates preserved when eligible prior values are carried forward.
Adaptive BioAge uses all three domains. KDM and Levine PhenoAge remain separate in Clinical Long View.
A controlled scientific core
Deterministic science.
Adaptive personalization.
Current core. Adaptive BioAge is computed from prespecified, version-controlled model rules rather than a self-training model. For the same eligible inputs evaluated under the same model version, the scientific calculation remains reproducible and the result can be traced back through domains and contributors.
Future personalization. BioMIR’s roadmap keeps adaptive intelligence downstream of that authoritative calculation. On-device learning may eventually characterize individual baseline, variability, temporal response, and action-response patterns without silently rewriting the equations that produced the underlying BioMIR result.
See the architecture and roadmapA model output expressed in age-equivalent units. A way to put signals into context.
Adaptive BioAge is a model output expressed in age-equivalent units relative to chronological age. Δ-years provides a common signed scale: negative values represent an age-equivalent reward and positive values an age-equivalent penalty. It is not a direct measurement of years gained or lost.
Short-term movement can reflect recent behavior, functional physiology, available data, or model assumptions. It does not establish a change in lifespan.
Interpret the result alongside its contributors, Data Confidence, CMA Freshness, and trends—not as a verdict on your health.
From signals to understanding
See what contributes.
Choose what comes next.
Follow a change over time. Inspect its contributors. Understand the measurements behind it. Move from Trends to domain and biomarker detail with Dashboard and Trends sharing period and statistic selections.
Take a closer look- 01
Look beneath the headline
Explore biomarker-level contributions instead of stopping at a single summary estimate.
- 02
Put patterns in perspective
Read trends with Data Confidence and CMA Freshness. Data Confidence summarizes daily input support; CMA Freshness keeps the age of carried cardiometabolic observations visible.
- 03
Make space for action
Review prioritized actions and organize them as Suggested, Planned, or Completed.
An open, careful approach
Good questions
are part of the science.
01 / Methodology
How the models work
Understand reference context, age-equivalent outputs, assumptions, data support, and the distinction between a model output and a clinical conclusion.
Explore the science02 / Resources
Go deeper into the evidence
Explore scientific reading and discussion, keeping population evidence separate from individual observations.
Browse resourcesEvidence-grounded explainers
Ask better questions
of biological age.
Short, referenced guides on the scientific ideas behind age-like health scores, wearables, functional physiology, cardiometabolic context, and longitudinal interpretation.
Browse all InsightsBiological age fundamentals
Biological age vs. health age: what do these scores actually mean?
Age-like health scores can be useful summaries, but the number only makes sense in the context of the inputs, reference population, outcome, and model assumptions behind it.
Read the explainerLongitudinal interpretation
Why can a biological-age estimate change from day to day?
High-frequency age estimates are dynamic summaries. Short-term movement can be informative without being interpreted as irreversible aging or lifespan change.
Read the explainerDigital health landscape
Apple Health Age and BioMIR: similar inputs, different modeling questions
Apple has announced Health Age for a forthcoming Health app update, making model transparency and interpretation an increasingly important comparison.
Read the explainerChoose your path
Use it. Examine it.
Help validate it.
01
Use BioMIR
Download the current public release and explore Adaptive BioAge, contributors, trends, and clinical long-view models.
App Store ↗ 02Study the model
Review the architecture, model assumptions, evidence boundaries, missing-data logic, and validation status.
Methods ↗ 03Research with BioMIR
Explore collaboration priorities for longitudinal convergence, preventive engagement, reliability, interpretability, and prospective validation.
Research & collaboration ↗Help shape BioMIR
Bring your curiosity.
Explore the beta and share your experience. Thoughtful feedback helps inform what comes next.