Architecture

Layers that can be inspected, permissioned and switched off.

Each layer communicates through explicit APIs and explicit consent. Nothing reads your knowledge without a permission you can revoke.

L01

TrueSelf ID

Identity layer

Identity belongs to the person, not the platform.

Passkeys, OAuth & MFA
Device and session management
Intent Profile: goals, values, constraints
Consent management, export, deletion

L02

OmniCloud

Knowledge layer

A personal and organizational memory system.

Memory Timeline
Knowledge Vault (encrypted)
Context Engine
Consent Mirror

L03

GenesisGPT

Intelligence layer

From prompting information to composing possibilities.

Creator, Researcher & Architect modes
Philosopher, Strategist & Scientist modes
Guardian mode for risk review
Multimodal analysis and synthesis

L04

Genesis Studio

Creation layer

Idea → Concept → Prototype → Simulation → Evaluation → Build.

Infinite canvas with AI copilot
Scenario branching
Knowledge references inline
Versioning and collaboration

L05

FreeWill OS

Autonomy layer

Assist decisions without covertly steering them.

Autonomy Dashboard
Nudge Detector
Decision Journal
Moral Sandbox

L06

Revelation Analytics

Decision Intelligence layer

Scenarios with stated evidence, assumptions and uncertainty.

Future Map
Drift: stated goals vs. observed behaviour
Impact over impressions
Anomaly and risk detection

L07

Ethics Engine

Ethics layer

Every material AI action is scored against declared principles.

Explainable Alignment Score
Bias and privacy assessment
Human approval & escalation
Versioned policy and audit logs

L08

The Council

Governance layer

Ethicists, scientists, engineers and community representatives.

Proposal and model review
Ethical escalation path
Public decision records
Accountability reporting

L09

Divine API

Developer Platform layer

Every layer addressable through explicit, permissioned APIs.

REST + GraphQL + Webhooks
TypeScript & Python SDKs
Sandbox, keys, rate limits
AngelicMQ event streams

L10

Impact Ledger

Impact layer

Capital → Activity → Output → Outcome → Long-term impact.

Custom impact frameworks
Social & environmental measures
Outcome verification
Longitudinal reporting

L11

Observatory

Observatory layer

Civilization-scale signals, rendered honestly.

Innovation and risk trends
Knowledge network maps
Environmental indicators
Scenario models

L12

Foundation

Infrastructure & Security layer

Privacy designed in, never appended.

RBAC + ABAC, tenant isolation
Encryption in transit and at rest
Audit trails and anomaly detection
Observability and cost control

Agent orchestration

No agent has unlimited authority

Agent	Responsibility	Default autonomy
Research Agent	Finds and synthesizes knowledge	Suggest only
Strategy Agent	Evaluates opportunities and threats	Suggest only
Architect Agent	Designs technical systems	Draft + approval
Ethics Agent	Reviews potential harms and conflicts	Always on
Data Agent	Analyzes structured data	Scoped datasets
Creative Agent	Generates concepts and variants	Draft + approval
Project Agent	Tracks execution and dependencies	Read + notify
Governance Agent	Monitors policy requirements	Always on

Permissions, tools, budgets, data access and approval requirements are set per agent, per project.

Data layer

Six shapes of memory

Relational
Users, organizations, permissions, projects, transactions
Graph
Knowledge relationships across people, ideas and outcomes
Vector
Semantic search and retrieval for AI context
Time-series
Metrics, telemetry and impact measurement
Object
Documents, media, datasets and generated artifacts
Event store
Immutable system events streamed through AngelicMQ

AngelicMQ

Event-driven by default

USER_CREATEDPROJECT_CREATEDKNOWLEDGE_UPDATEDAI_ANALYSIS_COMPLETEDPOLICY_TRIGGEREDREVIEW_REQUESTEDDECISION_RECORDEDIMPACT_UPDATED

AI memory

Three tiers, all revocable

Short-term context
The current conversation or task. Discarded on completion.
Project memory
Scoped to one project, visible to its members, exportable.
Personal knowledge memory
Long-term context, only with explicit authorization.

Every memory carries source, timestamp, confidence, permission, retention policy and a delete control.

See the architecture resolve a real objective.

Run the demonstration