Memory-first
Known facts are retrieved before an LLM is invoked, reducing cost, latency, and unnecessary inference.
Sovereign AI systems
Keep your AI, memory, and operations under your control.
NeuNuc brings local-first inference, durable context, research workflows, and operator tools into one system—built for people who need useful AI without surrendering their working memory.
Local by default
Local inference and embeddings keep private operating context on hardware you control.
Research when needed
External research is a deliberate, cited workflow—not a replacement for your internal source of truth.
One ecosystem, clear layers.
NeuNuc combines a desktop control surface, a local AI runtime, a context engine, a policy-aware OS control plane, and an optional bare-metal foundation.
Operator SurfaceDesktop, TUI, dashboards, and public documentation.
Context EngineFact lookup, source maps, retrieval, and token-budgeted context.
Inference RuntimeLocal llama.cpp, ONNX embeddings, and routed external research.
Control PlanePlans, approvals, event history, replay, and GTM evidence modules.
Build with NeuNuc.
We are preparing the public release. Join the update list for system demos, documentation, and early-access information.