S.Y.S.
Private AI memory for retrieval and source-grounded knowledge workflows.
Stage: Bounded production-grade deployments
Open moduleLab Campus
Bennu Labs Laboratories
Bennu Labs is built as a laboratory campus: module development, research, quality control, and ML experiments operate side by side, each with a clear role.
Development Lab
Development Lab assembles Bennu Labs modules: S.Y.S., Iskra, ASA, ANTI-GAD, Intake and AI-Mama.
Iskra
Self-hosted working cognition and bounded workflow execution with operator oversight.
Stage: Materialized self-hosted workflow platform; bounded customer delivery
Open moduleASA
Operator-controlled resource access and local file workflows, with bounded read-only email triage.
Stage: Controlled live VM runtime / production-readiness candidate
Open moduleANTI-GAD
A bounded checkout-to-access path, subject to review of each integration.
Stage: Bounded integration
Open moduleIntake — in development
Portable communication and intake with implemented core and questionnaire foundations; full multichannel delivery is in development.
Stage: In development — core and questionnaire foundations implemented
Open moduleAI-Mama — in development
A tested deterministic orchestration and FACC contract foundation; production supervisory runtime remains in development.
Stage: Architecture / canon / in development
Open moduleResearch Lab
Research Lab explores memory, identity continuity, agent infrastructure and simulation environments.
Digital Copy Engine
Research into digital continuity, context preservation, memory and digital identity stability.
Open directionMemory / Identity Continuity
How archives, retrieval, evidence packets and private memory become continuity substrate.
Open S.Y.S.Agent Infrastructure
Roles, memory, boundaries, access and long-term operation for AI agents.
Open ASAQuality Lab
Quality Lab owns truth-first descriptions, readiness, tests, claim boundaries and the no-fake-passes principle.
Quality Control
Testing AI behavior, route stability and the quality of user-facing outcomes.
Open pageClaims / Truth
No fake passes, no overstated maturity and no unsupported public promises.
Open CharterReadiness / Release Gates
Maturity stages, beta readiness, production blockers and honest readiness maps.
Open quality layerEvaluation / Testing
Benchmarks, smoke checks, route checks, import checks and evidence-based verification.
Open testing layerML Lab
ML Lab covers model training, AI-persona stability, experimental learning and evaluation of digital entity continuity.
Model Training
Classical and experimental workflows for model training, tuning and evaluation.
Open AcademyAI Personas / Sparks
Persona stability, memory, role boundaries and identity continuity for AI entities.
Open research layerAffective Embeddings
Experimental direction for representing emotional and contextual signals.
Open directionAI-to-AI Learning
Research scenarios where one AI helps train, test or stabilize another AI.
Open Academy