Your second brain is on a server nobody else indexes

Move the folder to a container on your rented bare metal box. The files URL serves it; hoody-code edits it; a workspace keeps the editor, a file browser, and a search tab open on any device. Ripgrep runs server-side through the files grep endpoint. The file journal keeps every revision without a sync service.

Nothing is used for training. Hoody's API sees that the container exists; it doesn't see the markdown, because the proxy serving the URL runs on your server. Share one URL with your partner for the grocery list; keep the rest to yourself.

Your notes are files on a disk you rent, and the only search index that touches them is yours.