Search

rag-aio canonical domain models and Protocol contracts

rag-aio context engineering: token budgets, dedup, expansion, citations

rag-aio multi-level caching: memory, disk, sqlite backends

rag-aio: RAG facade, CLI, and service composition

Deterministic, equal-area coverage commitments: polygon → rHEALPix cells → 32-byte root, with openings and set algebra.

Run local LLMs on the hardware you already have: compatibility doctor and memory planner for llama.cpp on old NVIDIA GPUs and Apple Silicon.

Command line and Python client for System One — the registry for decision models

SecureChain CLI - the supply-chain gate, as a single static binary

Boruna: a deterministic, capability-safe programming language and framework for LLM-native applications