AUTH WORKSPACE
Open Sign InSign In
Confirm identity and preserve a safe return route. Identity does not create permission.
PLATFORM ARCHITECTURE
Govern the relationship between humans, intelligence, memory, and action.
The Human Code is designed as a provider-independent control plane. Human identity and authority remain primary; models and tools are selected as capabilities under policy.
PUBLIC WORKSPACE DIRECTORY
Know what each surface can actually do before you open it.
A page is not called “live” just because it exists. These labels separate working tasks, read-only views, previews, reference surfaces, identity flows, and protected access.
AUTH WORKSPACE
Create a Human PageCreate a Human Page
Start public-page onboarding, choose a permanent path-based username, and continue through the identity gate.
WORKSPACE
Open editorHuman Page Editor
Authenticated owners can edit and publish their public Human Page projection.
LOCAL WORKSPACE
Open MirrorContinuity Mirror
Create and clear a browser-local reflection. Reflection does not become authority.
WORKSPACE
Open Export VaultExport Vault
Prepare or register an export with visible success/error state. Live upload depends on authenticated configured storage.
SYNTHETIC WORKSPACE
Open workflowSmart Family Flow
Exercise need selection, claim, assignment, follow-up, resolution, and closure using synthetic demo state.
READ-ONLY
Review consent stateConsent Review
Inspect visible consent records. Revoke/narrow mutation is not connected on this page yet.
READ-ONLY
View app accessApp Access
Inspect current or preview app-access records without implying grant/revoke controls.
PREVIEW
View notification modelNotifications
Inspect the notification model. It is not represented as a shared live delivery inbox.
PLANNED
View connector planConnector Hub
Review planned OAuth/import connector boundaries. No connection is authorized or created here.
REFERENCE
Inspect WalletContinuity Wallet
Inspect identity, consent, route, receipt, packet, and audit concepts on a production-hosted public reference surface.
REFERENCE
Inspect agent architectureAgentic Ops
Inspect the governed agent operating model. This page does not execute agents.
INFORMATION
Open verification indexVerification Status
Understand source, CI, preview, production, and runtime proof. This is not live uptime telemetry.
INFORMATION
Open Trust CenterTrust Center
Review privacy, security, professional, and product boundaries.
RECOVERY WORKSPACE
Open Recovery ModeRecovery Mode
Reduce the experience to one safe next step and return to the Wallet sign-in flow.
PROTECTED
Open archive gatePrivate Technical Archive
Access-controlled technical showcase. The gate fails closed when private credentials are unavailable.
DETERMINISTIC LIFECYCLE
Make every transition explainable.
The deeper implementation preserves a longer governed lifecycle; this public view compresses it to the decisions enterprise teams need to understand.
Request→Classify→Resolve authority→Derive policy→Select capability→Execute→Verify→Receipt→Memory eligibility→Commit / Ephemeral / Discard
Capability may be delegated.Authority does not transfer implicitly.
Provider substitution must preserve continuity.Identity, authority, provenance, and durable memory remain independent of a single vendor.
Execution does not imply retention.Memory disposition is a separate governed decision.
CORE CONTRACT
Five primitives. One explainable routing record.
The public architecture follows the locked Human Code Global Intelligence Fabric boundary rather than creating another model gateway.
01
HCProvider
An eligible source of intelligence or execution capability.
02
HCCapability
The task the system actually needs: reason, search, code, extract, translate, act, or verify.
03
HCProviderPolicy
Privacy, sovereignty, retention, region, trust, cost, and authority compatibility constraints.
04
HCRoutingDecision
Why a provider was eligible, excluded, scored, selected, or held behind a human gate.
05
HCExecutionReceipt
A computable record of request, authority, policy, provider, evidence, result, state change, and memory disposition.
RETURN PATH
Human-owned continuity
Receipts and eligible state return to governed continuity rather than disappearing into a model session.
DESIGN INVERSION
Human → Authority → Policy → Capability → Intelligence.
NOT THE OLD MODEL
Model → Agent → Tools → User is not enough.
That sequence makes the model the center of gravity. Human Code makes the person, their authority, and the governing policy the center instead.
- Models remain replaceable.
- Tool visibility never creates permission.
- Agent consensus cannot self-authorize.
- Evidence remains distinct from generated explanation.
BOUNDARIES
Governance should fail closed.
The architecture reserves consequential side effects for explicit authority evaluation, approval where required, execution, independent verification, and receipt creation.
Context locality
Retrieve only the information needed for the task, subject, purpose, and classification.
Revocation propagation
Changes to permission should affect future retrieval, routing, caching, and retention.
Side-effect membrane
Messages, filings, purchases, deployments, and protected mutations require governed execution rather than conversational intent alone.