Deployment on Linux

This guide covers running Celly as a long-lived service on a Linux host with systemd --user. As on Windows, the bot must run as the user who owns the sbx daemon and its credentials.

1. Host bootstrap

Install sbx (Docker Sandboxes), then log in and initialize the network policy:
sbx login sbx policy init balanced sbx version
Celly targets sbx >= 0.45.0.

Install Node 24 with fnm (or nvm):

curl -fsSL https://fnm.vercel.app/install | bash
fnm install 24
fnm default 24

2. Build and configure

git clone https://github.com/LegendArtur/bot-celly.git ~/discordAI
cd ~/discordAI
npm ci
npm run build
cp .env.example .env   # fill in DISCORD_TOKEN and DISCORD_GUILD_ID
node dist/index.js     # first foreground boot; Ctrl-C once it is ready

3. systemd user service

Create ~/.config/systemd/user/celly.service:

[Unit]
Description=Celly Discord bot
After=network-online.target

[Service]
Type=simple
WorkingDirectory=%h/discordAI
ExecStart=/bin/bash -lc 'exec node dist/index.js'
Restart=on-failure
RestartSec=5
KillSignal=SIGTERM
TimeoutStopSec=30

[Install]
WantedBy=default.target

bash -lc loads your shell profile so fnm/nvm and sbx are on PATH; exec keeps Node as the main process so SIGTERM reaches Celly’s shutdown handler. Enable it and keep it running after logout:

systemctl --user daemon-reload
systemctl --user enable --now celly
loginctl enable-linger "$USER"
journalctl --user -u celly -f

4. Backups and log rotation

Celly backs up DATA_DIR/bot.db with SQLite VACUUM INTO on an interval and prunes old copies. It also rotates bot.log and every per-project server log.

Variable	Default	Behavior
`BACKUP_INTERVAL_HOURS`	`24`	Hours between backups; `0` disables backups.
`BACKUP_KEEP`	`7`	Backup files kept before the oldest is pruned.
`LOG_MAX_BYTES`	`5000000`	Rotate a log after this many bytes.
`LOG_MAX_FILES`	`3`	Rotated copies kept (`.1`…`.N`).

Backups are named data/backups/bot-<ISO>.db. To restore one, stop the service, replace the database, and start it again:

systemctl --user stop celly
cp data/backups/bot-2026-09-27T00-00-00-000Z.db data/bot.db
systemctl --user start celly

5. Admin page (optional)

Set ADMIN_PORT (default 4560, 0 disables) to serve a loopback-only status page and JSON API:

GET / — HTML status page
GET /api/projects, GET /api/health
POST /api/projects/<channelId>/start|stop
GET /api/logs/<channelId>?lines=200 — redacted log tail
GET /api/audit?limit=100 — when the audit log is enabled

It binds 127.0.0.1 only and has no authentication: reach it through an SSH tunnel (ssh -L 4560:127.0.0.1:4560 host) and never expose the port.

6. Verifying a deployment

sbx diagnose reports a healthy daemon and authentication.
systemctl --user status celly is active (running).
curl -s http://127.0.0.1:4560/api/health returns {"ok":true,...}.
data/bot.log shows the preflight passing and the Discord client logging in.

7. Troubleshooting

Symptom	Cause	Fix
`sbx CLI not found`	PATH does not include `sbx` in the unit	Keep `ExecStart=/bin/bash -lc …`.
Boot fails with an `sbx login` error	Expired credentials	Re-run `sbx login` as the service user, then `systemctl --user restart celly`.
Service stops at logout	No lingering user session	`loginctl enable-linger "$USER"`.
Admin port unreachable	It listens on loopback by design	Use an SSH tunnel; do not expose it publicly.

Get started

Guides

Reference

Project

Deployment on Linux

1. Host bootstrap

2. Build and configure

3. systemd user service

4. Backups and log rotation

5. Admin page (optional)

6. Verifying a deployment

7. Troubleshooting

​1. Host bootstrap

​2. Build and configure

​3. systemd user service

​4. Backups and log rotation

​5. Admin page (optional)

​6. Verifying a deployment

​7. Troubleshooting

1. Host bootstrap

2. Build and configure

3. systemd user service

4. Backups and log rotation

5. Admin page (optional)

6. Verifying a deployment

7. Troubleshooting