Pre-launch · joining the waitlist
Your server heals itself at 3am.
Nightwatch detects broken services, repairs them automatically, and rolls back bad deploys — before anyone notices. No alerts at 3am. No SSH sessions at 3am. No users complaining at 9am.
Early access · open-source agent · $9/mo (Pro)
✅ You're in! We'll email you when the beta opens.
⚠️ Something went wrong — try again or email us directly.
1
Install
curl -sSL install.sh | bash — 5-minute setup, runs as your user, not root.
2
Declare
Define healthchecks and repair rules in a simple YAML file. Your scripts, your permissions.
3
Sleep
If a service dies, Nightwatch repairs it. If a deploy breaks it, Nightwatch rolls it back. Audit log of everything.
🩺Real healthchecks — catches "alive but broken" (HTTP 500, crash-loops), not just dead processes.
🪜Repair ladder — restart → recreate → custom script → escalate with logs. Deterministic, every time.
↩️Auto-rollback deploys — service dies right after a deploy? Reverted automatically in 60s.
🕯️Dead man's switch — if the whole machine dies, you get alerted with context, not silence.
🔓Open-source agent — audit exactly what runs on your box. Runs offline, only heartbeats go out.
🤖No AI in the loop — deterministic rules only. If X happens, Y always runs. No surprises on your server.
90% of outages fix themselves in under 5 minutes.
The other 10% alert you with logs before anyone else knows. 🦉
The other 10% alert you with logs before anyone else knows. 🦉