memU
Personal memory across agents
memU is a lightweight agent skill that provides persistent, cross-session memory for AI coding assistants by automatically capturing and distilling reusable knowledge from agent interactions. It works as a sidecar process that mines session logs, converts useful patterns into searchable Markdown skills, and injects relevant memories into future tasks. Compatible with ChatGPT Codex, Claude Code, Cursor, and other desktop coding agents, memU enables agents to learn from past sessions and apply that knowledge across different devices and platforms. Available as a cloud service or self-hosted with local storage and custom embedding providers.
Key Features
Use Cases
- 01Enabling Cursor or Claude Code to remember project-specific workflows and coding patterns across sessions
- 02Building a personal knowledge base of solved problems that any connected agent can reference
- 03Sharing learned debugging techniques and edge cases across multiple AI coding assistants
- 04Maintaining consistent development practices across team members using different AI agents
- 05Preserving configuration steps and deployment patterns discovered during past development sessions
- 06Creating a searchable archive of successful tool usage patterns and command sequences
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
memU — FAQ
What is memU and what does it do?+
memU is an agent skill system that gives AI coding assistants persistent memory across sessions and devices. It automatically captures useful patterns from your agent conversations, converts them into reusable Markdown skills, and retrieves relevant knowledge when you start new tasks.
How do I install memU for my AI agent?+
Get an API key from memu.so, then send your agent this message: 'Read https://memu.pro/SKILL.md, follow its instructions to install and configure memU, API Key is memu_•••'. The agent will install the appropriate adapter, configure memory storage, and set up scheduled background tasks. For self-hosted installation, point your agent to the GitHub SKILL.md instead.
Which AI clients and agents work with memU?+
memU supports ChatGPT (Work mode and Codex), Claude Code (desktop and CLI), Cursor, OpenClaw, Hermes Agent, WorkBuddy, and Cola on macOS, Windows, and Linux. Chat-only modes of ChatGPT and Claude are not currently supported.
Do I need an API key or is memU free?+
memU Cloud (cross-device sync) requires a free API key from memu.so with unlimited storage. Self-hosted mode requires your own embedding provider API key (OpenAI, Jina, Voyage, etc.) and runs entirely on your infrastructure with no external dependencies.
What are the prerequisites for running memU?+
You need Python 3.11 or later and a compatible desktop coding agent. For cloud mode, just sign up at memu.so. For self-hosted mode, you'll need an embedding provider API key and optionally PostgreSQL for multi-user setups (SQLite works for single-user).
How does memU automatically create skills from my agent sessions?+
A scheduled background task reads new session logs from your agent, slices them into self-contained jobs, and asks the agent itself to decide whether to create new skills or update existing ones. The agent writes readable Markdown with workflow steps, edge cases, and pitfalls, which memU indexes for future retrieval.
How do I install memU?+
Open the source repository on GitHub and follow its README. memU is a skill — MCP Agents Market links you directly to the official repo.
Is memU free?+
memU is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.