How do I make my Obsidian vault the memory Claude Code, Codex, and my other tools read?

Point every tool at the vault where it already is, and have each one read it before it helps, with nothing moved or converted. That’s how my own system works, written down as a blueprint, alexandria., which your model reads and then builds on your computer around the vault you already have.

If Claude Code or Codex already reads your vault through a CLAUDE.md or an AGENTS.md, you already run the core of this, and what’s left is what one vault can’t do alone.

When your model finds a vault, a Logseq graph, or a folder of markdown, it points the system at it rather than copying it in, and writes where it lives into the one guide that Claude Code, Codex, and Cursor each read at the start of every session. Obsidian stays where you read and write your notes, and nothing in the vault is moved, converted, or replaced.

Then it adds what one vault lacks: a session where your model brings up what’s most worth thinking about in your notes, pushes back, and writes what you settle in your own words; a one-tap save from your phone, so what you find away from your desk reaches your files; your tools working from the same record, with your best model handing out the work; and, if you join, what the people in your life choose to share with you.

Every part is plain files on your own computer, yours to change or drop, and your model changes nothing until you say yes to its plan.

Show the blueprint to your model, and it’ll tell you what your vault is missing, if anything.