I build IWE β an open-source memory system for you and your AI agents. It turns a directory of markdown files into a knowledge graph: an LSP server brings search, refactoring, and autocomplete into VS Code, Neovim, Zed, and Helix, while the CLI and MCP server give AI agents structured access to the same notes you edit by hand.
Things I care about:
- Plain text you own β notes as
.mdfiles in git; no cloud, no database, no lock-in - Editor-native tools β IDE features for your knowledge, right where you write code
- AI memory without a vector DB β retrieval by structure, not similarity guessing
- Rust β IWE processes 20,000 markdown files in under a second





