THE OPS ENGINEER
Plain-English investigations and one-click fixes you approve.
→ Helmsman chatYou're the captain. The Helmsman runs the ship.
Self-host production Kubernetes without a dedicated ops team. Helmsman diagnoses issues, monitors your cluster, and applies the fixes you approve.
N E S W db-01 HDG 000°
SCANNING
THE CREW FOUR HANDS YOU CAN'T AFFORD
THE ON-CALL SRE
Auto-remediation within your guardrails, while you sleep.
→ In-cluster agentTHE MONITORING TEAM
Natural-language alerts. No PromQL, no Prometheus.
→ Smart alertsTHE PLATFORM TEAM
One-click catalog. TLS'd apps on your own hardware.
→ App catalog THE NAVIGATOR
A teammate, not a console.
Ask in plain English. The Helmsman reads your logs, events and dependencies like a senior SRE, finds the cause, and proposes the fix as a one-click button. You stay at the wheel.
- Plain-English investigation across logs, events & dependencies
- Every fix is a one-click button. Nothing runs without your say-so
Helmsman 47.6°N 122.3°W
Why is checkout returning 502s?
Ingress is healthy. checkout has
2/3 replicas NotReady — the readiness probe on :8080 is failing. Root cause: the new image points DB_HOST at a service that no longer exists.
Proposed fix
1 ACTION
kubectl set env deploy/checkout DB_HOST=postgres-primary
Approve
Dismiss
THE NIGHT WATCH
It watches while you're away.
An in-cluster agent stands the night watch, catching drift and remediating within the guardrails you set. You wake up to a quiet morning instead of an incident.
- Stays inside the guardrails you set: spend caps, a kill-switch, and more
- Wake to a clean changelog, not a 3am page
Night Watch
watching · 03:14 UTC
- ingress-nginx — 502 error spike detected
- tracing upstream — checkout pod not ready
- auto-remediating — restarted 2 unhealthy replicas
- error rate back to 0.00% — verified & logged
ACTIVE GUARDRAILS
spend cap $50/mo no deletes kill-switch replicas ≤ 5
THE LOOKOUT
Tell it what to worry about.
Describe what matters in plain English: “text me if the postgres database goes down.” Helmsman keeps watch on the horizon and pings you the moment it happens.
no PromQL no Prometheus no YAML
New alert DRAFT
text me if the postgres database goes down
PARSED SMS postgres-primary is down
compiled to a live watch — no PromQL written
Helmsman now
ALERT — postgres-primary is DOWN. I'm standing watch and already investigating the fix.
THE HARBOR
Your own private cloud.
The catalog turns “I want to self-host X” into a running, TLS’d app in one click. Heroku ergonomics, docked at your hardware instead of someone else’s cloud.
One-click deploy from the catalog
Automatic TLS, issued and renewed
Runs on your hardware, so your data stays yours
App Catalog 55 apps
I want to self-host nextcloud
Nextcloud
Gitea
Vaultwarden
Plausible
Immich
Postgres
Deployed · TLS issued
https://nextcloud.harbor.local — live in 41s
Take the helm.
Self-host like you have an ops team: the crew you couldn't afford, diagnosing, fixing, and standing watch around the clock.