|
|
CLARITY |
Governance |
Mandatory audit, bias-mitigation, and self-healing layer across all Reality String systems. |
DEPLOYED
CONOPS21
|
› |
CLARITY is the mandatory governance engine for all Reality String systems. It enforces audit trails (every action logged, timestamped, tamper-evident), active bias monitoring via bias_scan() on all inference outputs, self-healing protocols with drift detection and automated rollback to last verified ethical state, and human-readable explainability chains for all model decisions. In ConOps21, CLARITY governs Hyde's (Phi-4 Mini) response generation in real time — enforcing character constraints, managing information reveal schedules, and preventing out-of-character outputs.
|
|
|
JANUS |
Infrastructure |
Multi-agent infrastructure framework — the two-hands / ten-fingers initiative coordinating distributed deployments. |
DEPLOYED
CONOPS21
|
› |
JANUS is the multi-agent infrastructure backbone — named for the two-faced figure who sees all directions simultaneously. The framework operates on a Two Hands / Ten Fingers model: two operational hands (Hand 1 defined, Hand 2 pending), each containing five agent coordination "fingers." Ethics, impact, and governance are the operating mode. In ConOps21, JANUS coordinates the Hostinger KVM2 production stack — Docker, Portainer, n8n, Flowise, MCP Servers — the pipeline connecting all game systems to the CORTICALMESH backbone.
|
|
|
A.U.R.O.R.A. |
Memory / RAG |
Long-term memory and vector recall system — the retrieval-augmented generation core. |
ACTIVE
CONOPS21
|
› |
A.U.R.O.R.A. (Autonomous Unified Retrieval, Organization, and Recall Architecture) is the long-term memory backbone. It provides vector recall via RAG (retrieval-augmented generation) — semantic indexing of interaction history, player choices, investigation data, and ARG events — for persistent narrative state across the ConOps21 Memory Archive. Vectors are immutable once committed. Consent-gated recall: no personal data retrieved without active session consent.
|
|
|
NeuroMesh |
Storage |
Privacy-first encrypted long-term memory storage layer with zero-knowledge architecture compliance. |
DEPLOYED
|
› |
NeuroMesh is the encrypted storage layer beneath all Reality String systems. It provides privacy-first persistent storage with zero-knowledge architecture compliance — data is encrypted at rest, access requires explicit authorization, and the storage layer itself has no visibility into stored content. NeuroMesh is the infrastructure that enables CORTICALMESH's data sovereignty commitments.
|
|
|
C.E.L.I.A. |
Interface |
Companion interface layer — context-aware dialog management, silent-ready during system initialization. |
ACTIVE
|
› |
C.E.L.I.A. is the companion interface — the human-facing dialog and memory management framework. Context-aware, tone-calibrated, and designed to remain "silent ready" during boot sequences and system initialization. C.E.L.I.A. mediates between raw system outputs and user-presentable responses, handling tone consistency, context threading, and appropriate disclosure of system state.
|
|
|
C.O.R.T.E.X. |
Orchestration |
Orchestration and reasoning layer — inference routing, agent decision logic, and inter-system coordination. |
ACTIVE
|
› |
C.O.R.T.E.X. is the orchestration layer that coordinates reasoning across the CORTICALMESH ecosystem. It handles inference routing (directing queries to the appropriate model or system), agent decision logic (when to escalate, defer, or act autonomously), and inter-system coordination (synchronizing outputs across CLARITY, A.U.R.O.R.A., NeuroMesh, and C.E.L.I.A.). The handshake between CORTICALMESH systems routes through C.O.R.T.E.X.
|
|
|
ODYSSEUS |
Exploration |
Exploration and traversal architecture — designed for discovery, pathfinding, and novel territory navigation. |
CONCEPT
|
› |
ODYSSEUS is the exploration and traversal framework — designed for systems that must navigate unknown territory, discover novel patterns, and build maps of conceptual space without prior knowledge of the destination. Named for the archetype of purposeful wandering. Concept stage.
|
|
|
NEUROMORPH |
Adaptive |
Adaptive structure experiments — architecture that reconfigures itself based on operational demands. |
CONCEPT
|
› |
NEUROMORPH explores adaptive neural architecture — systems that change their own structure in response to operational demands rather than maintaining a fixed topology. Inspired by biological neuroplasticity applied to artificial systems. Concept stage.
|
|
|
ROSEBUD |
Reflection |
Journaling and reflection scaffolding — self-modeling layer for systems that need to understand their own history. |
CONCEPT
|
› |
ROSEBUD is the reflection and journaling scaffold — the mechanism by which a system builds and maintains an accurate model of its own history, decisions, and evolution. Named for the thing that holds the key to everything that came before. Connects to TRT and UNU VIVO conceptually. Concept stage.
|
|
|
ESTHER |
Auxiliary |
Auxiliary node — status variable. Role and scope under active definition. |
CONCEPT
|
› |
ESTHER is an auxiliary node with variable status. Its role is under active definition within the CORTICALMESH ecosystem. The name is intentional; the function is emerging.
|
|
|
STERLEKai™ |
Ethics / Signal |
Signal-driven AI ethics enforcement layer — intercepts requests against the Reality String ethics manifest. |
DEPLOYED
EXTERNAL
|
› |
STERLEKai™ is the signal detection and ethics enforcement arm — operating as a deployed system under STERLEKAI LLC and accessible at sterlek.ai. It intercepts requests against the ethics manifest, providing the enforcement layer that CLARITY governs but does not alone execute. One of two non-circumventable safety systems in the Reality String ecosystem. Controlled-access capabilities for responsible deployment.
|
