Developer Brad Morris built a local database system for AI memory, replacing scattered text files with structured storage

Brad Morris, a developer, created ra-h_os — an open-source project that stores AI conversation history and context in a local database instead of plain text files and markdown documents, making memory searchable and organized.

Instead of the AI forgetting context between sessions or searching through scattered files, the database lets the AI instantly retrieve relevant past conversations and facts — similar to how Gmail finds old emails rather than you manually reading through folders.

For people building personal AI assistants or chatbots, this means the AI can now maintain actual long-term memory without losing information, making it more useful for ongoing projects, research, or daily tasks where continuity matters.