This repository contains no application code. It is the local orchestration layer: a base Compose file with Postgres and the shared network, plus three overlays that stack on top of it.
up-dev.sh (accepting --monitoring), down.sh, the loud-failure reset-db.sh, and bootstrap-glitchtip.sh, which creates and reconciles the error tracker's organization, projects, uptime monitors, heartbeats, and the alert rules from code.
Until October 2026 this repository also ran production: a prod overlay pulling the published images behind Watchtower, an up-prod.sh, and a backup and restore-drill pair on host systemd timers. All of that moved to a single-node k3s cluster managed by Flux from the public be-infra repository, where deploys are commits, the backups and the weekly restore drill are CronJobs, and monitoring is a shared platform stack. What stays here is everything you need to run Beyou on your own machine.
The infrastructure and monitoring architecture topics document both sides, this local stack and the production cluster, from the laptop it all runs on to the alert that fires when the snapshot scheduler goes quiet.