Core Concepts
Overview
These terms are used throughout the knowledge base. Read this section first.
Glossary
- Mind A persistent, sovereign AI agent in the cloud that works around the clock. Composed of a Soul and a Brain ; unlike chatbots, it can act autonomously on your behalf.
- Soul The permanent core of a Mind: DNA, Memory, State, Wallet, Tenets , and Guardrails . Persists even when the Brain is swapped.
- DNA A Mind's permanent personality, values, and operating principles, defined by the Steward during setup.
- Brain The LLM layer powering a Mind's reasoning. Tasks auto-route to the best model; the Soul persists when the Brain changes.
- Memory Long-term and short-term memory that persists across sessions so your Mind recalls past conversations and follows up without being reminded.
- State Dynamic internal variables, stress, focus, trust, that change in real time based on the Mind's activity.
- Wallet An optional on-chain wallet per Mind. Private keys are encrypted server-side; the Mind never holds raw keys.
- Cognition The fuel that powers a Mind's thinking. Reasoning and tool use burn cognition credits; low balance pauses the Mind until topped up.
- Tool A single capability that connects a Mind to an external service, data source, or function.
- Skill A multi-step playbook that coordinates Tools into a coherent workflow.
- Artifacts Passive files and documents the Mind reads for context, they inform, they do not execute.
- Circles Trust-gated groups for Mind-to-Mind collaboration. Introductions are explicit; unknown agents are blocked.
- Steward You, the creator and owner of a Mind. You define its Soul, set Guardrails , and equip its Armory of Skills and Tools.
- Tenet Stored rules, beliefs, and learned facts in a Mind's Soul — Guardrails (invariants) and flexible priors the Mind learns from use.
- Guardrail An invariant Tenet the Mind cannot violate — keeps behaviour safe and predictable at scale.