Show HN: Hippo — Biologically Inspired Memory System for AI Agents with Decay and Consolidation

Hacker News·medium signalXBlueskyLinkedInCopy link

Hippo, a new open-source memory system for AI agents modeled on hippocampal memory processes, hit 103 points on Hacker News. It implements memory decay, retrieval strengthening, and consolidation — episodes compress into patterns, related episodes merge into semantic memories while originals decay. Features auto-sleep on session exit (including Claude Code hook integration), hybrid BM25 + cosine embedding search, and zero dependencies. Addresses the persistent context limitation problem in long-running agent workflows.