NevaMind-AI/memU

This page introduces an external open-source repository. It is not an HDATF product.

What it does

memU is an agent-driven memory system that keeps a personal LLM wiki shared across sessions, agents and devices. It captures session knowledge in scheduled background tasks, retrieves it into later tasks, and distills reusable skills from agent history.

How it helps ATF

Worth comparing with how Harness keeps memory for a task and how Company Brain connects records with work. Storing memory as a wiki and turning agent history into skills could be a reference for both.

License

Custom license Custom or non-standard license. Read the license file before any reuse.

More in this category

• volcengine/OpenViking
  OpenViking is a context database for AI agents that stores documents, user memories and skills as a virtual filesystem agents browse with commands such as ls, read and write. Layered summaries let agents load full content only when needed.
• thedotmack/claude-mem
  Claude-mem captures what an agent does during sessions, compresses it into semantic summaries with AI, and injects relevant context into later sessions. It works with Claude Code, Codex, Gemini, OpenCode and other agents.
• garrytan/gbrain
  GBrain gives AI agents a memory that several agents can share. It stores explicit facts with their sources, supports corrections and withdrawal, and starts with keyword retrieval, adding semantic search and background enrichment when needed.
• memvid/memvid
  Memvid packs data, embeddings, search structure and metadata into a single file that AI agents can search without a database server. Memory is written as append-only frames with timestamps and checksums, so past memory states can be queried.
• MemPalace/mempalace
  MemPalace stores AI conversation history as verbatim text, without summarizing, and retrieves it with semantic search. Its index organizes people, projects and topics so searches can be scoped, and data stays local unless the user opts in.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.