Graphnosis explores a novel approach to AI knowledge representation: dual-graph structures (directed + undirected edges over the same node set) serialized in .gai — the AI Knowledge Graph binary format, optimized for machine comprehension, not human readability. The architecture mirrors the brain: the similarity graph is the cortex (long-term storage), the graph builder is the hippocampus (indexing), and the query engine is the prefrontal cortex (retrieval). Graph construction costs $0 (pure JS, no embedding APIs).
GitHub|Created by Nelu Lazar|MIT License