HOME/SPECIFICATION/INTELLIGENCE

HUMANOS 03 · CORE INTELLIGENCE PLATFORM

Engineered for biological signal. Zero noise, zero streak anxiety.

Commercial wellness applications hoard dozens of disconnected graphs and guilt-trip you over vanity streaks. HumanOS executes a deterministic personal loop: ingesting continuous biometrics, filtering out 200+ possible interventions, and generating strictly 3 explainable actions backed by empirical proof.

TEST IN SIMULATOR →VIEW EVIDENCE VAULT ↗LAUNCH TODAY APP ↗
PILLAR 01State Engine
MULTI-STREAM TELEMETRY

Continuous Biological Compilation

Transforms fragmented wearable streams and subjective check-ins into a unified, calibrated physiological state vector.

Most wellness apps hoard 40 conflicting charts and overwhelm users with raw sensor noise. The HumanOS State Engine executes an idempotent compilation pipeline: reading raw biometric telemetry, applying verified normalization profiles, filtering stale observations, and compiling calibrated domain scores across sleep, circadian rhythm, energy, recovery, focus, and movement.

Supported IngestionOura Ring Gen 3, Whoop 4.0, Apple Health, Garmin Epix, CGM, Manual Logs
Normalization Policyevidence-normalization-v1.2 with unit & freshness bounds
Signal IntegrityFail-closed: unprofiled metrics cannot become implicit 0–100 scores
Compilation Latency< 25ms in-memory deterministic state compilation
PILLAR 02Action Engine
HUMANIQ · RULE OF 3

Deterministic Daily Planning

Filters out 200+ possible habits down to strictly three high-leverage actions tailored to current biological reserve.

Analysis paralysis destroys behavioral adherence. HumanIQ evaluates your actual biological state and goals, applies safety contraindications, and returns exactly three actions for today. If personal data confidence is below the threshold or the vertical is baselining, it prescribes MEASURE_FIRST rather than risking unwarranted interventions.

Action ConstraintMaximum 3 daily actions. Zero cognitive overload.
Causal MechanismEvery action is bound to an empirical biological mechanism
Safety TiersLOW, MODERATE, HIGH, EXPERT_REVIEW, BLOCKED with fail-closed gates
Operating ModesNORMAL, RECOVERY, TRAVEL, PERFORMANCE, FOCUS, SLEEP_PREP
PILLAR 03Knowledge Engine
DUAL-STREAM LINEAGE

Modern Science & Ancestral Crosswalk

Explicitly separates certified biomedical trials from 3,000-year ancestral protocols without declaring unverified equivalence.

HumanOS refuses to flatten human knowledge into a generic recommendation blob. Modern peer-reviewed randomized controlled trials (PubMed/DOI indexed) and historical ancestral manuscripts (witness-level philological tracking) are kept strictly separate. Claims carry transparent evidence states: ESTABLISHED, STRONGLY_SUPPORTED, EMERGING, TRADITIONAL, or REFUTED.

Canonical Sources62 catalogued sources across 4 independent shards
Epistemic Hierarchy7 rigorous tiers (E1 Anecdotal → E7 Meta-Analysis of Multicenter RCTs)
Contradiction Edge TraversalSurfaces counter-evidence and challenge relationships
Consumer Safety GateOnly CONSUMER_SAFE certified claims attach to action guidance
PILLAR 04Verification Engine
CAUSAL ATTRIBUTION

Meaningful Change, Not Vanity Streaks

Calculates Minimum Meaningful Change (MMC) and causal attribution confidence to prove what actually moved your biological baseline.

Duolingo-style streak gamification creates toxic anxiety and rewards compliance over health. The Verification Engine looks for real, durable physiological progress: verifying whether your protocol adherence caused the observed delta while keeping vital guardrails intact. If you take a day off to rest, the system logs it as recovery, not a broken streak.

Attribution ModelMathematical causal attribution confidence [0.00 – 1.00]
Meaningful Change GateDelta must exceed measurement noise threshold (MMC)
Guardrail IntegrityAutomatic disqualification if vital biomarkers degrade
Anti-Streak GuaranteeZero guilt, zero vanity streaks, zero behavioral addiction

SYSTEM TOPOLOGY & DATA BUS

The Deterministic Personal Intelligence Pipeline

Observations flow inward across strict privacy boundaries, compile into sovereign state vectors, pass through snapshot-bound epistemic filters, and emit bounded actions without opaque third-party cloud leakage.

01 Ingest

Wearables & Check-ins

Oura, Whoop, Apple Watch, and manual subjective logs ingested idempotently.

02 Compile

State Engine

Applies metric profiles, freshness windows, and computes domain confidence.

03 Retrieve

Evidence Graph

Resolves claims through snapshot-bound, certified consumer EvidencePacks.

04 Decide

HumanIQ Engine

Filters 200+ habits down to strictly 3 daily actions with explicit physiological rationales.

05 Verify

Attribution Engine

Verifies minimum meaningful change (MMC) and confirms guardrail integrity.

CROSS-REFERENCE

The Research Vault

Examine the 62 canonical sources, evidence datasets v2, and the epistemic classification standards underpinning HumanIQ decisions.

Explore Research Vault →
SIMULATION SANDBOX

Interactive Telemetry Simulator

Manipulate sleep hours, resting heart rate, and wearable streams to test how the HumanOS State and Action Engines respond.

Launch Full Simulator →