Schedule automatic ingest

Read this if you want OMem to keep itself up to date — ingesting new mail, files, and events on a schedule instead of you running omem ingest by hand. OMem uses macOS’s own launchd timer; there’s no background daemon to manage.

Install the schedule

You already chose an interval during setup, so just run:

omem install

This reads the interval from your config and writes a launchd job that runs omem ingest on that schedule. Because re-runs are cheap (OMem only touches what changed), a frequent interval is fine.

To use a different interval than what’s in your config — or if you skipped that step — pass --schedule (it overrides the config value just for this install):

omem install --schedule 15m     # 3m, 15m, 1h, … — overrides config

It's a short-lived job, not a daemon

There’s no always-running process. launchd wakes a brief omem ingest, it does its work, and it exits. Nothing sits resident in memory between runs — that’s design principle P1, and it’s why OMem is “zero ops.”

Check it’s healthy

omem install status

↻ replay

$

This command has its own status codes (0 = healthy, 1 = not installed, 2 = not loaded, 3 = installed but the last run failed or nothing ran in 24h) — see exit codes. It’s a one-line health probe you can poll from a monitoring script.

Uninstall

omem install --uninstall

Removes the launchd job. Your wiki and data are untouched — only the schedule stops.

Two things worth knowing

It uses the globally-installed omem. The scheduled job must point at the same omem binary you grant Full Disk Access to — the one on your PATH (from brew or the curl install), not a development copy. This keeps the macOS permission identity consistent so scheduled mail ingest doesn’t silently fail.

File-change triggering is opt-in. By default the job runs purely on the timer. You can add --watch-roots to also trigger when files in your roots change, but it’s off by default: ingest is a multi-minute batch, not a real-time sync, so the timer is the recommended pattern.

omem install --watch-roots   # also trigger on file changes

What’s next

• Observe ingest — watch what the scheduled runs are doing.
• The ingest lifecycle — why frequent runs stay cheap.