mnemon-dev

mnemon-dev / mnemon

Public

LLM-supervised persistent memory for AI agents — graph-based recall, cross-session knowledge, single binary. Works with Claude Code, OpenClaw, and any CLI agent.

github.commnemon-devmnemon#readme agent-framework agent-memory ai-agent ai-tools claude
18
0
100% credibility
Found Feb 23, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

Mnemon adds persistent, graph-based memory to AI agents, letting them retain and recall insights across sessions under LLM supervision.

How It Works

1
💡 Discover Mnemon

You hear about a simple tool that gives your AI chat buddy a lasting memory, so it never forgets important details from past talks.

2
📥 Grab it quickly

With one easy command, you add Mnemon to your computer, ready to enhance your AI helper.

3
🔧 Connect once

Run a single setup to link it smoothly to your favorite AI chat app, like Claude or similar.

4
🧠 Memory awakens

Now, as you chat with your AI, it automatically saves key facts, choices, and links between ideas in its own growing knowledge web.

5
🔍 Recall anytime

Your AI pulls up relevant memories on its own, or you can peek at the collection to refresh old thoughts.

🚀 Smarter companion

Over time, your AI becomes a true partner, remembering everything and helping you make better decisions effortlessly.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 18 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 mnemon?

Mnemon is a Go single binary delivering LLM-supervised persistent memory for AI agents, tackling the issue of LLMs forgetting cross-session knowledge due to context limits. It builds a graph-based store with temporal, entity, causal, and semantic edges for intent-aware recall, using your Claude subscription as the supervisor—no extra API keys. One `mnemon setup` integrates with Claude Code or OpenClaw via hooks/plugins, plus CLI commands like `recall`, `remember`, `gc`, and `viz` for manual mnemonic management.

Why is it gaining traction?

Stands out by making your host LLM (Claude, CLI agents) the judge for what to store or forget, avoiding embedded models' overhead. Graph-based recall fuses keywords, optional Ollama vectors, and traversals for smarter results than vector-only tools. Zero-config persistence across sessions, with deduping, decay, and multi-store isolation hooks developers building stateful agents.

Who should use this?

Developers crafting Claude Code or OpenClaw agents needing mnemonics for cross-session knowledge, like iterative code gen where early decisions vanish. Ideal for CLI agent users in long analysis chains or prototyping graph-based agent memory.

Verdict

Worth trying for agent builders seeking lightweight, LLM-driven persistence—solid docs, tests, Go binary. But 10 stars and 1% credibility signal early stage; production users should monitor maturity.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.