Show HN: Oubli – Persistent fractal memory for Claude Code (github.com)

Oubli, a newly announced memory system for Claude Code, offers a robust solution for persistent memory management in AI interactions. This tool facilitates the seamless storage and retrieval of a user’s memories and insights, allowing for a more personalized and contextual AI experience. Users can easily export their information from platforms like ChatGPT, Gemini, or Claude.ai, and import it into Oubli. The system employs a layered memory architecture that distinguishes between core memories, synthesized insights, and raw memories, ensuring that essential information about the user is always accessible and maintained across sessions. Significantly, Oubli utilizes a hybrid search mechanism that combines traditional keyword searches with semantic embeddings, allowing for nuanced retrieval of information based on concept associations rather than mere keyword matches. Features such as proactive memory updates, which automatically save new information without user prompts, and a visual memory graph that organizes memories hierarchically, enhance user engagement and insight generation. The local-first design ensures data privacy, as all user information is stored on their own machine, marking a notable advancement in personalized AI memory capabilities. This tool not only enriches user interactions but also sets a new benchmark in AI's ability to remember and leverage past conversations for improved future interactions.