Catalog of AI Knowledge Retrieval, Memory and RAG Systems (github.com)

A new reference catalog has been introduced that maps AI knowledge retrieval, memory, and Retrieval-Augmented Generation (RAG) systems, tackling the fragmented landscape of over 100 projects with overlapping functionalities. The catalog emphasizes the parallels between human memory mechanisms and the architectures employed in AI systems, such as vector databases for associative recall and knowledge graphs for relational reasoning. This initiative aims to clarify project functionalities, stack positions, hardware requirements, and cognitive mappings, ultimately assisting those building or assessing AI knowledge infrastructures. For the AI/ML community, this catalog is significant as it addresses the confusion caused by numerous rapidly evolving projects with unclear boundaries. By systematically organizing these systems from low-level infrastructure to high-level cognitive functions, the catalog provides invaluable guidance on evaluating and developing AI stacks. The inclusion of a cognition mapping table relating back to human memory processes enhances understanding of each project’s intent and applicability, supporting better-informed decisions in local-first setup implementations, where hardware compatibility is critical.