Basic Memory is an open-source memory system that lets AI assistants actually remember your work across conversations. Published on GitHub, it has gathered about 4,300 stars and is under active development in 2026. Instead of re-explaining your project to your AI every session, you and the assistant build a persistent knowledge graph together, stored as plain Markdown files on your own computer and connected through the Model Context Protocol (MCP).
The tool is local-first and privacy-first: notes live in ordinary Markdown files you can read, edit, and back up with tools you already use, including Obsidian. When you chat with an assistant such as Claude, ChatGPT, Cursor, or Codex, Basic Memory gives it structured access to that knowledge โ entities, observations, and relations โ so context survives across sessions and even across different AI tools. A hosted cloud edition removes the local setup entirely and adds shared team workspaces where people and their assistants read and write the same knowledge base in real time.
Basic Memory follows a freemium model. The open-source version is free to self-host with all core features. The optional cloud service starts at $15 per month with a 7-day free trial, and a Teams edition adds shared workspaces for organizations.
Basic Memory is for developers, researchers, and knowledge workers who use AI assistants daily and are tired of repeating themselves โ anyone who wants a durable, portable memory layer that belongs to them rather than to a single AI vendor.