whereismyai
Hy Memory
Hy Memory is a production-grade dual-system cognitive memory engine and Python SDK for self-evolving LLM agents.
Hy Memory is a production-grade dual-system cognitive memory engine and Python SDK named hy-memory for self-evolving LLM agents. It offers specialized memory pipelines across three task scenes including hierarchical long-term cognitive memory for chat, two-layer fact and reusable-experience memory for working, and version-evolution diff-chain tracking for coding. MemHub is built on top of Hy Memory as a local-first desktop and CLI product that unifies conversation history and coding traces from AI applications such as Claude Code, Cursor, Codex, OpenCode, and Qwen Code into a single managed runtime. The system supports vector stores like ChromaDB, Qdrant, and FAISS. It operates through lite, pro, and ultra modes with varying levels of LLM extraction and graph database support.
Hy Memory at a glance
Key features
- Hierarchical long-term cognitive memory for chat
- Two-layer fact and reusable-experience memory for working
- Version-evolution diff-chain tracking for coding
- Automatic scene routing for writes and searches
- Source isolation with application filtering
- System 2 schema induction and cross-domain sweeper
- Per-call embedding endpoint override
Use it for
- Manage long-term personal chat memory and preferences
- Store session facts and reusable experiences for productivity work
- Track software engineering code changes and diff chains
Alternatives to Hy Memory
- AI Memory — Shared memory for coding agents that works across machines and team members.
- IHMT MEMORY — It shares one local memory across multiple AI agents without cloud storage or databases.
- claude-mem — It shares persistent, searchable agent memory across sessions and devices.
- Engram — Single Go binary with zero dependencies and built-in full-text search.
- Memori — Reduces LLM token usage and spend by over 97% with targeted recall.
- DocuMemory — Turns messy AI chat transcripts into clean, reusable context for future sessions.
- IAI Memory Engine — Provides persistent, local, and encrypted memory for AI agents without APIs.
- Supermemory — Integrates long-term memory into AI coding agents with a single prompt.
- M-Flow — It achieves a 93.7 percent top-one hit rate using a conical memory architecture.
- SimpleMem — Provides semantic lossless compression to reduce token consumption and maintain long-term context.
- Tigerless Agent Memory — Maintains agent history in portable, greppable Markdown files without API keys.
- MemOS — It delivers ultra-persistent memory, hybrid-retrieval, and token savings.
All Hy Memory alternatives