Show HN: Kore – local AI memory layer with Ebbinghaus forgetting curve (github.com)

Kore, a newly announced local AI memory layer, aims to revolutionize how AI agents retain information by mirroring human-like memory processes. Unlike existing solutions that require constant cloud connectivity and substantial reliance on large language models (LLMs), Kore operates fully offline. It employs the Ebbinghaus forgetting curve to manage memory decay, ensuring that only essential information lasts, while less critical notes fade quickly—transforming how AI agents prioritize and recall data without external dependencies. This innovation is significant for the AI/ML community as it enhances data privacy and operational efficiency. With features like auto-importance scoring, semantic search in over 50 languages, and intelligent memory compression, Kore allows for a leaner, more responsive database. Its design supports multi-agent systems where each agent only accesses its own memories, enabling secure and isolated interactions. Integrated with OpenClaw, a personal AI agent platform, Kore facilitates rich, context-aware experiences by automatically scoring and retrieving memories based on relevance and semantic similarity, promoting a more intuitive user experience.