Quick start
A fresh install has nothing to recall. These steps fill it, and these habits make recall better.
Give it a head start
Run these once in your project.
Terminal
claude-db flush
claude-db scan
claude-db seed --from-git| Command | What you get |
|---|---|
claude-db flush | Your past chats for this project, loaded into memory |
claude-db scan | A map of your code. It takes about a second and costs no tokens |
claude-db seed --from-git | Your commit history, loaded into memory so search has something to match |
Claude Code keeps chat files for about 30 days, so run flush early. After a session or two, older chats are also turned into short facts in the background, newest first, within a daily limit.
You can build the map from inside Claude Code instead, with the /cdb-scan skill.
Work in ways that help
- Be specific. "Make the order feed reconnect with backoff" finds memory. "ok" and "continue" do not.
- Say why. "We chose a websocket because polling hammered the API" is kept as a decision with its reason.
- State a rule once. Tell Claude to remember it, or run
claude-db remember "always use pnpm in this repo". - Commit finished work. Uncommitted work stays listed as open.
- Wrap private text in
<private>...</private>. It is removed before saving. - Use
/catchupwhen you return and/handoffwhen you stop. See Skills.
Ask something
From here you do not run commands. Ask Claude what you would normally ask:
what breaks if I change closeObservations?
why does capture read the transcript instead of the diff?
what was I in the middle of yesterday?The first is answered by the code graph, the second by memory, and the third by /catchup.
Look at what it holds
Terminal
claude-db search <words>
claude-db stats
claude-db viewview opens this project's memory in the browser, and stops when you press Ctrl+C.