/ Dragon AI

Your server has a guardian.

Not a chatbot that points at menus an agent that actually runs your server: it reads the telemetry, finds the cause, and fixes it with one tap.

Live telemetry Safe 1-tap fixes Everything reversible· Live now on every server in your dashboard

/ the shift

The difference is doing, not telling.

Every other host

A chatbot that points at menus

“To whitelist a player, go to Settings → Access.” You still do all the work and still don’t know why it lagged.

Dragon

An agent that runs your server

Reads live metrics and logs, pinpoints the cause with real numbers, proposes a concrete fix and applies it safely once you approve.

/ what Dragon does

From diagnosis to done.

live

Diagnosis with evidence

“Why did my server lag at 21:00?” → an answer with the TPS dip, the cause per plugin/entity, and the log lines to prove it.

live

Safe 1-tap fixes

Dragon proposes a concrete change as a diff. You tap approve; it applies behind a snapshot.

live

Everything reversible

Every change is snapshotted first. Don’t like it? One tap and your server is back the way it was.

live

Proactive monitoring

Dragon spots a lag spike or a looming problem and flags it in your dashboard and Discord before your players notice.

/ how it feels

Just ask.

survival-smp · minecraft telemetry live
Why did my server lag at 21:00 yesterday?

Dragon

Between 20:53 and 21:17 your TPS dropped from 20 to 14.2. 83% of tick time went to entity processing; VillagerTradeRebalance accounted for 31%, and one farm held 1,487 entities.

Propose a fixShow the evidence

/ your data stays yours

Your data stays yours.

Dragon works entirely within your own server and account — performance, logs, config, backups and billing — and uses that to give you concrete answers right away. Your data stays private and inside your own environment: it’s never shared and never used to train AI models. You stay in control.

  • Works only within your own server & account
  • Your data is never shared or sold
  • Never used to train AI models

/ safe by design

An agent with keys needs brakes.

The model proposes, code disposes

The AI never gets write access. It only emits a typed plan; plain, auditable code executes it.

You approve every change

Nothing changes without your explicit tap. Approval is per plan never remembered, never batched.

Snapshot before every action

Every change is captured first. Undo restores the exact prior state.

Fully audited

Who, what, when, which diff and which rollback point it’s all in your audit log.

/ where it’s going

This is only the beginning.

Now

Diagnosis & evidence

Live telemetry, logs and incidents ask Dragon why something happened.

Now

1-tap fixes

Config changes and restarts, safe behind snapshot + approval.

Soon

Safe upgrades

Snapshot → clone → update → compatibility scan → test → report, before your live server is touched.

Later

Proactive guardian

Dragon handles small things within your limits and hands you the incident report over coffee.

/ ready to start

Every GameUnit server gets Dragon.

No add-on, no surcharge for the diagnosis. Pick a server and open the Dragon tab.

See the pricing Included with every server