OpenKnowledge
AI-native WYSIWYG document editor and local knowledge management solution.
OpenKnowledge, officially launched on June 25, 2026, by Inkeep (YC W23), is an open-source, AI-native WYSIWYG document editor and local knowledge management (LLM Wiki) solution designed for a modern development environment where humans and AI coexist. While traditional document editing tools focus solely on visual convenience for humans, this tool acts as a bridge, enabling AI agents and LLM-based knowledge retrieval systems to perfectly navigate and edit document structures. Technically, it is developed based on proven editor cores such as Tiptap, Prosemirror, and CodeMirror, and incorporates a powerful engine that allows for visual editing of local Markdown and MDX files without compromising the original content.
Previously, the conventional method of document creation involved Markdown files scattered in local directories or isolated in cloud-based services, making it extremely difficult for AI coding assistants or automation agents to directly read and modify them, hindering the development process. OpenKnowledge directly addresses these limitations by functioning like a "library where AI assistants can enter and organize books in real-time." It features the Model Context Protocol (MCP) as a core component, providing an AI-friendly ecosystem that allows agents such as Claude Code, Cursor, and Codex to directly query, update, and manage versions of the knowledge base in the local file system without user intervention.
The value of utilizing this tool in the biotechnology and bio-infrastructure environment is even more pronounced. By enabling OpenKnowledge in an Obsidian directory containing thousands of gene analysis results or clinical trial data protocols, AI agents can directly analyze the documents through MCP, automatically updating key gene marker information or synchronizing related paper abstracts to keep the knowledge base up-to-date. Researchers can instantly visually review the final report through a sophisticated Notion-style layout without directly handling Markdown code, and this entire process is tracked through Git, maximizing the reliability and reproducibility of clinical documentation.
💻 System Requirements
0 (CPU-based operation; when performing local embedding and AI operations, utilize VRAM from an external API or a local agent environment)
Approximately 500MB (includes application installation space and local cache; document storage capacity is separate)
⚡ Installation
4-1. Quick Start
Download the macOS Desktop App:
Visit the official website and download the .dmg file for macOS, then drag and drop it into your Applications folder.
CLI Installation and Project Initialization (Linux, Windows, Intel Mac):
npm install -g @inkeep/open-knowledge
cd your-project
ok init
4-2. Detailed Installation
-
Prerequisites: Ensure that Node.js (version 24 or higher) is installed in your local environment.
node -v -
Global Package Installation: Use npm to install the OpenKnowledge CLI in the global environment.
npm install -g @inkeep/open-knowledge -
Specify and Initialize Project Folder: Navigate to the working directory where your Markdown files will be stored, create a project, and scaffold the AI agent integration files.
cd /path/to/your/knowledge-base ok init -
Run the Web Editor: Open a local server and launch the editor in your web browser.
ok start --open
FAQ
What is OpenKnowledge?
OpenKnowledge, officially launched on June 25, 2026, by Inkeep (YC W23), is an open-source, AI-native WYSIWYG document editor and local knowledge management (LLM Wiki) solution designed for a modern development environment where humans and AI coexist. While traditional document editing tools focus solely on visual convenience for humans, this tool acts as a bridge, enabling AI agents and LLM-based knowledge retrieval systems to perfectly navigate and edit document structures. Technically, it is developed based on proven editor cores such as Tiptap, Prosemirror, and CodeMirror, and incorporates a powerful engine that allows for visual editing of local Markdown and MDX files without compromising the original content. Previously, the conventional method of document creation involved Markdown files scattered in local directories or isolated in cloud-based services, making it extremely difficult for AI coding assistants or automation agents to directly read and modify them, hindering the development process. OpenKnowledge directly addresses these limitations by functioning like a "library where AI assistants can enter and organize books in real-time." It features the Model Context Protocol (MCP) as a core component, providing an AI-friendly ecosystem that allows agents such as Claude Code, Cursor, and Codex to directly query, update, and manage versions of the knowledge base in the local file system without user intervention. The value of utilizing this tool in the biotechnology and bio-infrastructure environment is even more pronounced. By enabling OpenKnowledge in an Obsidian directory containing thousands of gene analysis results or clinical trial data protocols, AI agents can directly analyze the documents through MCP, automatically updating key gene marker information or synchronizing related paper abstracts to keep the knowledge base up-to-date. Researchers can instantly visually review the final report through a sophisticated Notion-style layout without directly handling Markdown code, and this entire process is tracked through Git, maximizing the reliability and reproducibility of clinical documentation.
When should I use OpenKnowledge?
AI-native WYSIWYG document editor and local knowledge management solution.
📝 Update Notes
No update notes yet.
🧪 Related Code of Life
No related Code of Life posts yet.