open-knowledge: a WYSIWYG markdown editor with collaborative AI-editing and MCP support for LLM wikis
open-knowledge: a WYSIWYG markdown editor with collaborative AI-editing and MCP support for LLM wikis
What it solves
OpenKnowledge provides a private, local, and free markdown editor designed for creating knowledge bases, LLM wikis, and engineering specs. It bridges the gap between a simple text editor and a collaborative AI workspace, allowing users to manage documentation that is easily accessible to both humans and AI agents.
How it works
It functions as a WYSIWYG editor that can be opened on any local folder containing markdown or mdx files (such as Obsidian vaults or existing codebases). It integrates with AI harnesses like Claude, Codex, and Cursor via MCP (Model Context Protocol) and CLI, enabling collaborative AI-editing. Under the hood, it uses git and GitHub for team sharing and automatic synchronization.
Who it’s for
It is built for developers and teams who need a "Notion meets VSCode" experience for maintaining technical documentation, agent second brains, and knowledge graphs while keeping their data local.
Highlights
- AI-Native Editing: Direct integration with Claude, Codex, Cursor, and OpenCode for collaborative drafting.
- Agentic Capabilities: Built-in support for MCP, skills, and agentic search for LLM wikis.
- True WYSIWYG: A visual editing experience similar to Google Docs or Notion for markdown files.
- Flexible Deployment: Available as a native macOS app or a local web app for Linux and Windows via CLI.
- Developer Tooling: Includes a built-in TUI (Terminal User Interface) and support for embeddable HTML and rich components.
Sources
- undefinedinkeep/open-knowledge