Skip to content

AI Systems

GNUS.ai is both a distributed execution platform and an application platform for modular AI. It should not be described only as distributed RAG, sharded LLM inference, or generic agents running on GPUs.

  • GCS (Genius Cognitive System) governs the cognitive request: planning, Semantic Core reasoning, routed Expert Language Models (ELMs), grounding, tools, structured GAML memory, and answer checks.
  • SuperGenius / GNUS handles eligible distributed processing, network participation, ledger and settlement work. It does not replace the GCS controller, model routing, or memory governor.
  • Applications, including Genius AI Boss, expose these capabilities to users. GCS may execute locally or on a private network without sending each request to a public swarm.

The pages below cover infrastructure components, including retrieval, storage, pub/sub and retraining design. They are not a complete implementation-status inventory of GCS. See the GCS architecture documentation for the cognitive specifications and the Master Platform Architecture for component ownership.

Release status: SuperGenius mainnet implementation is complete as of October 2026; its public launch is held for GCS and initial application readiness. See Platform Status.