Self-hosted, privacy-first knowledge workspace with block-level references, Markdown WYSIWYG editing, and AI agent collaboration.
- Stars
- 1,779
- Forks
- 106
- Open issues
- 1
- Last commit
- 2 days ago
What is it?
IWE is a local-first knowledge base built on plain Markdown files. Notes link to each other and can belong to several topics, so ideas stay connected across a project. It is aimed at people who want a shared memory that both they and their AI agents can read, edit, and organize.
It ships in four forms that operate on the same files: a native Mac app called iWe, a language server that adds link autocomplete, backlinks, search, and safe renaming to Neovim, Helix, VS Code, and Zed, a command-line tool for searching and managing notes, and an MCP server that lets Claude, Codex, and other MCP-compatible agents work with the notes. The project does not include an AI model; agents connect through the user's own setup.
Because everything is stored as Markdown in a folder on the local machine, notes can be backed up, synced, or kept in git without an IWE cloud account. Agent-saved notes are ordinary files the user can review or delete.
Key features
- 01Markdown notes that link to each other and can belong to multiple topics.
- 02Language server with link autocomplete, backlinks, search, and safe renaming for Neovim, Helix, VS Code, and Zed.
- 03Command-line tool for finding and managing notes, including lexical search.
- 04MCP server so Claude, Codex, and other compatible agents can read and write notes.
- 05Local Markdown storage with no required cloud account.
Use cases
- Keeping project specs, plans, and decisions in one place that an AI agent can also reference.
- Writing and linking notes inside an existing editor setup without switching tools.
- Managing a Markdown notes folder from the terminal or in scripts.
- Giving coding or marketing agents persistent, reviewable memory across sessions.
Quick start
- 1
Install IWE with Homebrew
brew install iwe-org/iwe/iwe - 2
Initialize IWE in your notes folder
iwe init - 3
Search your notes by content
iwe find --lexical "project decisions"
Things to know
- Connecting an AI agent requires that agent's own setup and account.
- The README does not offer a Windows or Linux desktop app; the Mac app is the native GUI option.
Topics
FAQ
- What is iwe?
- Markdown knowledge graph with an LSP for your editor and a CLI and MCP server so AI agents can read and write the same notes. IWE is a local-first knowledge base built on plain Markdown files. Notes link to each other and can belong to several topics, so ideas stay connected across a project. It is aimed at people who want a shared memory that both they and their AI agents can read, edit, and organize.
- Is iwe free to use? What license does it have?
- iwe is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
- Is iwe still maintained?
- Actively maintained. The last commit to iwe was on Oct 18, 2026.
- What can I use iwe for?
- Keeping project specs, plans, and decisions in one place that an AI agent can also reference. Writing and linking notes inside an existing editor setup without switching tools. Managing a Markdown notes folder from the terminal or in scripts. Giving coding or marketing agents persistent, reviewable memory across sessions.
- How do I install iwe?
- Install IWE with Homebrew: brew install iwe-org/iwe/iwe Initialize IWE in your notes folder: iwe init Search your notes by content: iwe find --lexical "project decisions"
- Where can I get iwe?
- The source code and installation instructions are on GitHub: https://github.com/iwe-org/iwe
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An open-source, local-first knowledge base that merges docs, whiteboard canvas and tables in one workspace, as an alternative to Notion and Miro.
Open-source, local-first 3D building editor built on React Three Fiber and WebGPU, usable in the browser or from a CLI with MCP tools for AI agents.
Open-source notes and knowledge base app with a three-pane layout, instant multi-device sync, and built-in MCP for AI agents, runnable on Cloudflare or Docker.
MCP server and CLI letting AI agents reverse engineer native binaries, Electron apps, .NET assemblies and websites.
A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps.