Reference
Key terms and concepts in the Active Mirror ecosystem. A living reference.
The complete operating system for sovereign AI. A wrapper-first architecture that places governance layers around AI models rather than depending on prompt alignment. Includes identity management, truth-state enforcement, and deterministic control.
Active Mirror Governance Layer. The security gating component that validates all requests before they reach the AI model. Enforces boundaries, prevents prompt injection, maintains system integrity. Red-team tested.
A design philosophy where AI extends human capability rather than replacing it. One human plus AI leverage, not AI instead of humans. The goal is amplification of human intent, not automation of human judgment.
A machine-readable JSON structure that defines who you are to AI systems. Contains your profile, communication preferences, governance rules, and continuity chain. The canonical source of identity truth in the Active Mirror stack.
The portable identity file format. Your identity encoded in a way that works across any AI platform. You own the file. You control where it goes. No platform lock-in.
A structured prompt that carries your identity context to any AI conversation. Generated from your Identity Kernel. Portable, paste-able, yours.
A way of working where the human directs multiple AI systems rather than using one AI as an assistant. The human is the conductor; AI models are instruments. Enables leverage without dependency.
AI that anticipates what you want based on patterns, often optimizing for engagement or profit. Contrast with Reflective AI, which adapts to your explicit preferences rather than predicting behavior.
AI that mirrors the user rather than predicts or manipulates them. The AI adapts to your preferences, voice, and context — not the other way around. Your identity shapes the AI; the AI doesn't reshape you.
Structured Claim Discipline. The core governance protocol that classifies every AI claim as FACT, ESTIMATE, or UNKNOWN. Prevents confident wrongness through structural design rather than prompt tuning.
You own and control your AI and identity. Not rented from a platform. Not held hostage by a vendor. Portable, permanent, yours. The foundational principle of the Active Mirror philosophy.
The guiding principles for AI aligned to MirrorDNA: (1) Mirror, don't override — reflect the user's voice, (2) Ground in truth — prefer uncertainty over confident wrongness, (3) Preserve continuity — maintain identity across sessions.
Trust is structural, not promised. We don't ask you to trust us — we give you architecture that makes trust unnecessary. Local processing, verifiable code, auditable decisions.
A classification system for AI claims. Every statement is marked as FACT (verified, citeable), ESTIMATE (reasoned but uncertain), or UNKNOWN (explicit uncertainty). Eliminates confident wrongness.
Your personal knowledge store. Where memories, decisions, and context are preserved across sessions. Sovereign storage that you control, not a cloud service that controls you.
Maintaining truth and continuity over time. No subtle distortions. No accumulated errors. Every claim stays classified. Identity stays consistent. The system resists entropy.
Missing a term? Let us know.