Now open source

Your knowledge, co-authored by AI in real time

An agent-native knowledge platform where humans and AI co-create. VS Code meets Notion — connected to any AI agent via MCP.

Try it

Two modes, one source of truth

Switch between rich editing and raw markdown. Edit in either mode — the content stays in sync, just like the real product.

getting-started.md

Get started

Up and running in under a minute

Three steps from zero to a live, AI-collaborative knowledge base.

01

Run one command

npx @inkeep/open-knowledge

02

Open the editor and start writing

A rich WYSIWYG editor opens in your browser. Toggle to source mode anytime. Your content is saved as plain markdown files — no database, just files in a folder.

03

Connect an AI agent and collaborate in real time

Point any MCP-compatible agent — Claude, Cursor, Codex — at the server. The agent reads, writes, and searches your knowledge base alongside you, live.

terminal
$npx @inkeep/open-knowledge
Scaffolding .open-knowledge/ ...
Registering MCP server in .mcp.json ...
Server running at http://localhost:5173
MCP server ready for agent connections
Watching ./content for changes...
Found 12 markdown files in content/
Shadow repo initialized at .git/open-knowledge/
Ready — open http://localhost:5173 in your browser
open-knowledge — Getting Started
getting-started.md
open-knowledge — API Reference
Y
AI
2 connected
Live
You
AI Agent
typing...

How it works

From raw notes to structured knowledge

Inspired by Andrej Karpathy's LLM knowledge base workflow — automated and real-time.

01

You write

Rich editor or raw markdown. Real-time CRDT sync — your cursor never fights the AI.

02

AI co-authors

An agent writes alongside you — expanding, referencing, structuring — all live.

03

Knowledge compiles

A background agent organizes your notes into a cross-linked, structured wiki.

04

Query & grow

Ask questions against your knowledge base. Answers get filed back automatically.

“You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.”

Start building your knowledge base

One command to get started. Connect your favorite AI agent and let the knowledge compile itself.