CALYREN companion continuum formed from woven silver filaments, permissioned signal constellations and a protective open arc.

CALYREN™ · inside Project HECARYON™

Intelligence that communicates its limits.

The concurrent intelligence partner inside Project HECARYON™—designed to correlate human, garment, environmental and operational context while keeping provenance, confidence, uncertainty, permissions and human oversight visible.

Read the interaction contract↓Return to HECARYON→

RoleConcurrent partner / interface

ProgramInside Project HECARYON™

Current stageControlled local, synthetic-only Alpha baseline

ControlQualified human oversight

Explore by system layer

Move from overview to controlled detail.

Each route has one job, a distinct evidence boundary and a clear relationship to the wider Heartstone system.

Governance

Authority ladder, privacy-preserving learning and explicit non-authorizations.

Open route →02

Industry pathways

How CALYREN must adapt under sector-specific Domain Profiles.

Open route →03

Milestones

What must be evidenced before capability advances.

Open route →

What CALYREN is intended to be

A disciplined interface between signals and decisions.

CALYREN is conceived to help organize relevant inputs, surface system conditions, explain how an interpretation was formed and support a qualified person’s next decision within defined bounds.

What CALYREN is not

Not a separate product lineCALYREN belongs within the HECARYON program hierarchy.

Not an authorityIt should not replace qualified judgment, consent, accountability or required professional oversight.

Not an operational systemThe approved Alpha is local, synthetic-only, inactive or mocked. It has no live garment data, provider activation, persistent personal memory, autonomous tool use or production deployment.

Not a validated capabilityThe current interface and imagery are conceptual, not a performance demonstration.

Capability architecture

Make the reasoning conditions inspectable.

01 / Receive

Governed input

Accept only defined inputs with source, time, version, permissions and quality conditions attached.

02 / Interpret

Bounded context

Relate inputs to intended use, requirements, known conditions and explicit exclusions.

03 / Communicate

Confidence + uncertainty

Separate observation, inference, recommendation and unknowns in language a person can examine.

04 / Escalate

Human oversight

Defer, warn or route the matter when evidence, authority or operating conditions are insufficient.

Four-domain context architecture

Separate the sources before relating the meaning.

CALYREN should never collapse “inside” and “outside” into one ambiguous stream. Each domain carries its own source, quality, time, permissions and limitations. A Continuum Correlation Layer may then examine how they relate without turning an association into a diagnosis or command.

01 · Human State Domain

The person

Authorized physiological modalities, wearer reports, communication, accessibility preferences and individual baselines.

02 · Garment State Domain

The physical interface

Material-zone condition, fit context, microclimate, connection state, signal quality, serviceability and possible integrity change.

03 · Environmental Context Domain

The surrounding conditions

Temperature, location, air, water, facility or other context only from authorized instruments suitable for that use.

04 · Operational Context Domain

The task and authority

Role, mission, workflow, team procedure, equipment state, escalation route and rules governing the situation.

Continuum Correlation Layer

Relate trends across domains; preserve source and uncertainty; distinguish measurement, observation, correlation, inference and recommendation; support the wearer and authorized humans without replacing accountable judgment.

Personalization without loss of control

One CALYREN architecture. A bounded companion for each authorized user.

A CALYREN Companion Instance may learn how an individual prefers information to be timed, explained and delivered. It may establish permissioned personal baselines and accessibility settings. It must not rewrite safety rules, evidence thresholds, role authority or data rights.

CALYREN CoreThe controlled shared architecture, safety kernel, terminology, evaluation and release rules.

Companion InstanceThe user-bound interface, preferences, consent state and local context.

Domain ProfileQualified knowledge, modalities and boundaries for a defined sector such as rehabilitation or controlled-environment agriculture.

Mission ProfileTemporary role, task, environment, authority and escalation constraints for a specific activity.

Continuum Learning NetworkPrivacy-preserving, reviewed propagation of generalized learning—not unrestricted sharing of individual histories.

Modality & adaptability

The interface adapts to the person, context and consequence.

Future modalities may include visual, text, voice, haptic or structured system-to-system outputs. Availability does not equal authority: each channel needs its own consent, accessibility, privacy, interruption, failure and fallback contract.

Input modalities

Receive with source control.

Garment modules, approved external systems, wearer input and authorized professional context remain distinguishable.

Quality and failure states required
Live modalities need separate authorization

Output modalities

Communicate without hiding uncertainty.

Text, visual, voice and haptic pathways should express urgency, confidence and required human review appropriately.

Voice must support interruption and text fallback
No output may imply unsupported authority

Personal adaptation

Learn presentation—not permission.

Tone, detail, timing, language and accessibility may adapt to the user and setting.

Consent and revocation remain visible
Safety and evidence thresholds remain fixed

Operational adaptation

Use the right profile for the task.

Domain and mission profiles constrain vocabulary, inputs, escalation, latency and role-specific information.

Least-privilege information access
Clear inactive, degraded and unavailable states

Privacy-preserving federated continuum intelligence

Learn across the system without making a person’s raw life the shared dataset.

01 · Local

Personal data stays bounded.

Raw user histories, sensitive signals and preferences remain local or in explicitly authorized encrypted stores where possible.

02 · Minimize

Extract only what is needed.

Candidate learning is permissioned, purpose-limited, de-identified or transformed and screened for sensitivity.

03 · Review

CALYREN Core does not self-authorize.

Evidence, privacy, safety, bias, security and version review determine whether generalized learning may enter the controlled baseline.

04 · Version

Every change has identity.

Approved learning carries provenance, evaluation, rollback and affected-domain records.

05 · Distribute

Only qualified updates propagate.

Companion instances receive applicable improvements without receiving another user’s raw record.

Emergency access, if ever developed, must be minimal, role-bound, time-limited, purpose-specific and audited. It should disclose only the information needed for the authorized response and must not create unrestricted institutional access.

The interaction contract

Every response should show what it knows—and why it may be wrong.

This is a design requirement for future validation, not a claim that the interface currently performs these functions.

ObservationWhat input or system condition was detected or supplied?

ProvenanceWhere did it come from, when, under which version and permissions?

InterpretationWhich requirements, rules or models were applied?

ConfidenceHow strongly is the interpretation supported in this context?

UncertaintyWhat is missing, conflicting, estimated or outside scope?

Next actionWhat can a qualified person consider, verify, defer or escalate?

Illustrative dialogue logic

A voice should never hide the evidence boundary.

The following is a conceptual communication pattern—not a live system response or validated product demonstration.

System signal

“A condition has changed.”

CALYREN should identify the source, timing and relevance of the signal before interpreting it.

Bounded interpretation

“This may matter because…”

State the rule or requirement applied and separate an inference from a confirmed observation.

Uncertainty

“I cannot determine…”

Expose missing inputs, conflicting evidence and conditions outside the intended-use boundary.

Human decision

“A qualified reviewer should…”

Offer bounded options or escalation—not autonomous consequential action.

No medical, clinical, protective, safety-critical, emergency or autonomous consequential capability is implied. Any future use in a consequential context would require defined intended use, appropriate validation, qualified oversight and applicable regulatory consideration.

Collaborate on responsible intelligence→Review the evidence standard→

Inside the current build

CALYREN has a controlled foundation and a larger proposed architecture.

The public record separates implemented repository controls from architecture under review and capability that is not active.

Implemented baseline

Inert provider-shaped operation

Mock transport, inactive adapters, live-register controls, schema checks and fail-closed behavior.

Architecture under review

Core + Companion + Profiles

CALYREN Core, individualized Companion Instance, Domain Profile, Mission Profile and Continuum Learning Network.

Next proof

Canonical domain contracts

Source, time, unit, quality, provenance, uncertainty, purpose, authorization and retention envelopes.

Not active

Personal or consequential operation

No live provider, persistent memory, raw human-data learning, emergency access or autonomous consequential authority.

Review the development record↗

CALYREN directory