m

memU

Personal memory across agents

14.5k
Stars
+1208
Stars/month
133
Commits (90d)
1
Releases (6m)

Star Growth

+3.6k (33.3%)estimated from velocity
10.7k12.7k14.8kJul 2Sep 30

Overview

memU is a lightweight, agent-driven memory system that provides a shared LLM wiki across sessions, agents, and devices. It automatically distills reusable skills from agent history and works with various coding agents like Codex, Claude Code, Cursor, and OpenClaw.

Deep Analysis

Key Differentiator

Core memory logic is only 500 lines, making it compact enough to inspect, understand, and adapt.

⚡ Capabilities

  • • Memorize session knowledge via background tasks
  • • Retrieve relevant memory for future tasks
  • • Store memory as a reusable wiki

🔗 Integrations

ChatGPT (Work mode)Claude CodeCursorOpenClawHermes AgentWorkBuddy

✓ Best For

  • ✓ Users running multiple AI coding agents
  • ✓ Teams needing shared knowledge across agents
  • ✓ Maintaining consistent memory across development sessions

✗ Not Ideal For

  • ✗ End-user chatbots or generic AI apps
  • ✗ Non-AI development tools
  • ✗ Single-session or single-agent use cases

⚠ Known Limitations

  • ⚠ Chat mode not supported for some agents
  • ⚠ Retrieve support unverified for OpenClaw
  • ⚠ Requires specific model selection for Claude Code setup

Alternatives

See all 8 memU alternatives →

Works with memU

Tools that integrate with memU, often used together in the same stack.

Compare memU

Maintain memU?

Show your live rank in your README, or put memU in front of every visitor to AgentoolRank.