yoniassia

yoniassia / memclawz

Public

Three-Speed Memory for OpenClaw Agents — QMD working memory + Zvec hybrid search + auto-compaction

ai-agents hnsw memory openclaw qmd
28
5
100% credibility
Found Feb 28, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

memclawz provides OpenClaw AI agents with a local, multi-layered memory system offering instant access to recent tasks, rapid hybrid searches, and smart long-term recall without needing internet or external services.

How It Works

1
🧠 Discover better memory for your AI helper

While using your OpenClaw AI agent, you learn about memclawz, a simple upgrade that gives it three speeds of remembering things perfectly and super fast, all on your own computer.

2
🚀 Add it with one easy command

From your OpenClaw folder, you run a single ready-made script that sets everything up automatically without any hassle.

3
✨ It grabs all your past memories

In just minutes, it pulls in every bit of your agent's history, creates a quick notepad for current work, and starts a helper service to keep everything fresh.

4
📝 Update your agent's daily plan

Copy a short set of instructions into your agent's guide so it knows to check its new fast memory first for instant recall.

5
🔍 Test the speedy recall

Ask your agent about past tasks or details, and watch it find answers in milliseconds instead of slow searches.

🎉 Your agent remembers like a pro

Now your AI helper resumes work instantly, never loses track of ongoing tasks, and searches its full history blazing fast, making everything smoother and smarter.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 28 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 memclawz?

memclawz delivers three-speed memory for OpenClaw agents: QMD for instant working memory via JSON files, Zvec for hybrid vector-plus-keyword search, and Mem0 for smart long-term storage with auto-compaction. Built in Python, it runs 100% locally with no API keys, upgrading OpenClaw's built-in memory_search to handle hot tasks in under 1ms, indexed memories in 8ms, and historical recall in 100ms. Agents get persistent context across sessions via one-command install and endpoints like POST /search on localhost:4011.

Why is it gaining traction?

It crushes OpenClaw's 1.7s semantic-only searches with hybrid search that nails keywords and vectors, plus auto-indexing of new memories every 60s and QMD working memory that survives restarts. One script imports your full history, starts servers, and verifies setup—no config hell. Developers love the 50x speedups and zero-maintenance compaction that keeps logs lean.

Who should use this?

OpenClaw agent builders frustrated by session amnesia and slow memory lookups during tasks like research or trading workflows. Ideal for solo devs or teams running local agent fleets needing shared memory across specialists. Skip if you're not on OpenClaw or prefer cloud vector DBs.

Verdict

Worth cloning for OpenClaw users—solid docs, 34/34 passing tests, and MIT license make it production-ready despite 12 stars and 1.0% credibility score. Early but battle-tested; pair with the context optimizer script for sub-agent setups.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.