A knowledge base you run
Find the note you meant to keep.
Save a source, a thought, or a useful fix. Give it structure. Search the title and content when the question returns, then revisit how the entry changed.
See the workflowREFERENCE · SAVED LOCALLY
How the cache
keeps its promise
Keep the source close to the conclusion. When the same problem returns, search for the terms you actually remember.
01 / THE METHOD
Capture is only the beginning.
Brainstash is most useful when the detail you saved can be found, opened, and revised later.
- 01
Keep the useful detail.
Create an entry with a title and content. Add a source URL, category, or tags when they help you return to it.
- 02
Search what you saved.
SQLite FTS5 searches entry titles and content. Open a result to read the original note and its source.
- 03
Revisit its history.
Browse earlier title and content snapshots, and restore a previous version when you need it.
02 / CURRENT SCOPE
A source-run library, with a clear boundary.
Brainstash runs as a Bun API and SvelteKit web app with a local SQLite database by default. Set an AUTH_SECRET, start the two services, and create an account on that instance. There is no verified hosted Brainstash account service.
Media files are currently served by ID without authentication. Search filters run after FTS pagination, so filtered counts and pages can be incomplete. The illustrated note above explains the workflow; it is not captured product UI.
Repository access may be restricted. The project notes list the current setup requirements and limitations.