Memori is an open-source memory engine that addresses the stateless nature of AI agents by providing persistent memory capabilities. The system uses a multi-agent architecture with three memory modes (Conscious, Auto, and Combined) to intelligently manage context and reduce token costs. Built with SQL-first design supporting SQLite, PostgreSQL, and MySQL, it integrates with popular frameworks like LangChain, CrewAI, and Agno. The engine enables agents to remember past interactions, maintain consistency across sessions, and avoid repeating work, making them more efficient and human-like in their responses.