Skip to content

Add memsearch — markdown-first semantic memory for AI agents#15

Open
zc277584121 wants to merge 1 commit into
topoteretes:mainfrom
zc277584121:add-memsearch
Open

Add memsearch — markdown-first semantic memory for AI agents#15
zc277584121 wants to merge 1 commit into
topoteretes:mainfrom
zc277584121:add-memsearch

Conversation

@zc277584121
Copy link
Copy Markdown

New Entry

Field Value
Name memsearch
URL https://zilliztech.github.io/memsearch/
Open / Close Open source
GitHub URL https://github.com/zilliztech/memsearch
Category Memory Tool
Storage Vector

What is memsearch?

memsearch is a semantic memory search library for AI agents where markdown files are the source of truth and the vector store (Milvus) is a derived index — rebuildable anytime.

Key features:

  • Markdown-first — memories are plain .md files (human-readable, git-friendly, zero vendor lock-in)
  • Content-hash dedup — SHA-256 hash as primary key; unchanged chunks are never re-embedded
  • Hybrid search — dense embedding + BM25 sparse search with RRF fusion via Milvus
  • 5 embedding providers — OpenAI, Gemini, Voyage, Ollama, sentence-transformers (including fully local options)
  • 3 Milvus backends — Milvus Lite (embedded), Milvus Server, Zilliz Cloud
  • Claude Code plugin — a ready-made hooks-based plugin that gives Claude Code automatic persistent memory across sessions
  • Framework integrations — LangChain, LangGraph, LlamaIndex, CrewAI

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@zc277584121
Copy link
Copy Markdown
Author

Hi @Vasilije1990 @hande-k, could you please review this when you have a moment? Thank you!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant