iwe-org/iwe

Markdown knowledge graph — LSP for your editor, CLI + MCP memory for your AI agents

What it solves

IWE provides a way to turn a local directory of Markdown files into a queryable knowledge graph. It allows users to maintain full ownership of their notes in plain text while providing the structured, database-like retrieval capabilities needed for both human browsing in an editor and AI agents querying via a command line or API.

How it works

IWE treats Markdown files as a connected structure using two types of links: "inclusion links" for nesting (creating a hierarchy/tree) and cross-references for creating a web of relationships. It allows a single note to have multiple parents, meaning it can belong to multiple topics without duplicating files.

For AI agents, it provides a CLI and an MCP (Model Context Protocol) server that allows agents to search, retrieve, and refactor notes. It ensures safety through "expect" guards that validate the blast radius of mutations and document schemas that validate frontmatter and structure. It is built in Rust for high performance, capable of processing 20,000 files in under a second.

Who it’s for

  • Knowledge workers who use Markdown-based note-taking systems and want IDE-like features (search, refactoring, autocomplete) in their editors (VS Code, Neovim, Zed, Helix).
  • AI agent developers who want to provide their agents with a long-term, human-readable memory system that is version-controlled via Git and avoids the "similarity guessing" of vector databases.
  • Users of AI coding agents (like Claude Code) who want a structured memory for their project context.

Highlights

  • Plain Markdown ownership: No proprietary databases or cloud lock-in; everything is stored as .md files.
  • Model Context Protocol (MCP) support: Native integration with AI tools like Claude Desktop, Cursor, and Windsurf.
  • Graph-based retrieval: Retrieves notes with their parent context and children in a single call, rather than relying on simple similarity search.
  • LSP integration: Provides IDE-like features for multiple editors via the Language Server Protocol.
  • Guarded mutations: AI agent writes are validated against schemas and scope guards to prevent accidental corruption of the knowledge graph.

Related

  • Project
    agentscope-ai/ReMe

    A local-first, self-evolving personal knowledge base for AI agents that stores memory as human-readable Markdown files for long-term knowledge evolution.

  • Project
    text2future/flowix

    A Markdown notebook that turns local notes into durable working memory for AI agents via an MCP server and CLI.

  • Project
    aayoawoyemi/Ori-Mnemos

    An open-source persistent memory infrastructure for AI agents that uses a local knowledge graph of markdown files to implement cognitive-inspired retrieval and learning.

  • Project
    akitaonrails/ai-memory

    ai‑memory is a Rust‑based server that records every observation a coding‑assistant makes (prompts, tool calls, session boundaries) as plain markdown files in a git‑backed wiki. The server can be run locally, in Docker/Podman, or on a homelab box and is designed to be shared across many agents (Claude Code, Codex, Cursor, Gemini CLI, etc.), machines and team members. Hooks posted by agents are consolidated into readable wiki pages at session end; the next session receives a concise handoff so it can continue exactly where the previous one stopped. Storage is human‑readable markdown plus a SQLite index (FTS5, entity and optional vector search). By default it needs no LLM or embedding keys; optional Anthropic/OpenAI integrations add LLM‑generated summaries and semantic search. Multi‑user authentication, audit logs and per‑person attribution are built in. Install via Arch AUR, Docker wrapper, or native binary, start the server, register a client and its hooks, and the system works transparently.

  • Project
    thecodacus/understory

    A self-wiring, plain-markdown memory layer for AI agents that organizes facts into a cross-linked knowledge graph based on the Open Knowledge Format.