zilliztech

zilliztech / memsearch

Public

A Markdown-first memory system, a standalone library for any AI agent. Inspired by OpenClaw.

zilliztech.github.iomemsearch agent agent-memory claude-code claude-code-plugin clawdbot
676
63
100% credibility
Found Feb 11, 2026 at 30 stars 23x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

memsearch is a Python library for semantic search over markdown files to provide persistent memory for AI agents, compatible with various embedding providers and Milvus vector storage, including a plugin for Claude Code.

How It Works

1
🔍 Discover memsearch

You hear about a handy tool that gives your AI chats a long-term memory using everyday text notes.

2
📥 Get it ready

You set it up on your computer in moments, pointing it to a folder for your notes.

3
🧠 Feed it your notes

It reads your text files, breaks them into smart pieces, and remembers the important bits.

4
💬 Chat smarter

When you talk to your AI like Claude, it pulls up the best matching memories to help out.

5
👀 Watch it update

It keeps an eye on your notes folder and automatically adds new stuff as you go.

6
📝 Summarize sessions

It condenses long chats into short, useful notes for even better recall later.

✨ AI remembers forever

Your AI now has a perfect memory across all sessions, feeling like a trusted longtime friend.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 30 to 676 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is memsearch?

memsearch rips OpenClaw's markdown-first memory system into a standalone Python library for any AI agent. It indexes your markdown notes or logs into a Milvus-backed vector store for semantic search, with live file watching, smart dedup via content hashing, and LLM-powered compaction to compress old memories. Developers get a CLI for quick index/search/watch/compact commands and a simple async Python API to plug persistent memory into custom agents.

Why is it gaining traction?

It brings OpenClaw-inspired memory everywhere—no monorepo lock-in—with markdown as the git-friendly source of truth and a derived vector index you can rebuild anytime. The Claude Code plugin delivers zero-config persistent memory across sessions via hooks, auto-injecting relevant context on every prompt. Pluggable embeddings (OpenAI, local, Ollama) and Milvus scaling (lite file to Zilliz Cloud) make it flexible without vendor ties.

Who should use this?

AI agent builders needing semantic recall over conversation logs or docs, especially OpenClaw fans wanting it standalone. Claude Code users tired of session amnesia will love the drop-in plugin for automatic memory injection and drill-down. Python devs prototyping RAG pipelines over markdown knowledge bases.

Verdict

Early alpha with 17 stars and 1.0% credibility score—docs shine and PyPI install works, but expect rough edges in production. Grab it if you need OpenClaw-style agent memory now; otherwise watch for maturity.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.