iwe-org/iwe

What it solves

IWE addresses the problem of fragmented AI memory, where AI agents lose context between sessions and cannot easily access or share a user's existing project notes. It provides a persistent, human-readable, and agent-accessible memory layer that prevents users from having to repeatedly explain project details and decisions to their AI tools.

How it works

IWE functions as a shared workspace consisting of a folder of connected Markdown documents. It acts as a bridge between the user and their AI agents (such as Claude or Codex) using the Model Context Protocol (MCP). Users can interact with these notes through a native Mac app, their preferred code editor (like VS Code or Neovim), a command-line interface (CLI), or directly through an AI agent. Because the memory is stored as local Markdown files, the user retains full control, and the notes can be version-controlled or backed up like any other project file.

Who it’s for

This tool is for developers and power users who want to maintain a persistent, structured, and searchable project knowledge base that their AI agents can read, write to, and organize alongside them.

Highlights

• Shared context: Allows both humans and AI agents to read and update the same set of project notes.
• Human-readable memory: All stored information is in standard Markdown, ensuring users can always verify, edit, or delete what the AI has saved.
• Tool-agnostic: Supports integration with various AI agents via the Model Context Protocol (MCP) and works with standard text editors.
• Local-first: Notes are stored as local files, ensuring privacy and portability without requiring a proprietary cloud account.