SmartNote: Superação das Limitações das Ferramentas de Gestão de Conhecimento Pessoal mediante Ligações Semânticas Automáticas e Inteligência Artificial Local

Personal Knowledge Management (PKM) tools, such as Obsidian or Notion, prove useful for organising notes and information; however, they present significant limitations that hinder the discovery of implicit relationships, the expansion of cognitive capabilities, and the preservation of data sovereignty. Their near-exclusive reliance on explicit, manually created links tends to generate knowledge fragmentation and information overload, particularly as the volume of data increases. Furthermore, the use of cloud-based services introduces vulnerabilities that may compromise privacy and reduce the user’s effective control over their own information. In this paper, we present a personal knowledge management system designed to mitigate these limitations through the integration of natural language processing techniques, vector indexing, and local artificial intelligence. The developed solution, named SmartNote, employs semantic embeddings to identify contextual connections, uses the FAISS library for efficient vector indexing, and integrates a Retrieval-Augmented Generation (RAG) AI assistant, fully operational offline via Ollama. The results confirm the technical feasibility of this approach and point towards a new direction for personal knowledge management systems, offering enhanced contextualisation, scalability, and privacy, while providing users with control over their own information.