A migration looks scary.
Clone production, run it there, throw the clone away.
OPEN-SOURCE DEPLOYMENT. YOUR SERVERS. OPEN BETA.
Production-perfect
environments.
On your servers.
In seconds.
Clone any environment, data included. Deploy it to hardware you own. Roll back when it goes sideways.
Start deploying
curl -fsSL https://ployz.sh | sh- secondsto clone an environment, data included
- 0Kubernetes
- 1bill, and it's your server's
production staging review-42 · cloned from prod review-38 · shares staging db logs · metrics agent: promoting web…- Node
- Bun
- Python
- Go
- Rust
- PHP
- Ruby
- .NET
- Postgres
- MySQL
- Redis
- anything with a Dockerfile
OWN YOUR SERVERSCLONE EVERYTHINGREAD THE PLANLET THE CREW SAIL
CLONE
Clone full production. In seconds.
Everything production has. Nothing it doesn't. ZFS underneath, so a terabyte clones as fast as a megabyte.
| Feature | Capability | Note |
|---|---|---|
| Environment clone | Seconds, data included | vs 30s to 24h elsewhere |
| Rollback | Any snapshot, seconds | every deploy is undoable |
| Clone storage | 0 GB until you write | copy-on-write, ZFS underneath |
| New service on live data | One rope on the canvas | staging or production database |
BUILD
Drop a box. Get a service.
Canvas, Compose, CLI or an agent. One model underneath, your call on the way in.
ployz deploy -f compose.yamlployz host bootstrap # add a server> promote review-42 → production
awaiting your approval…THE CREW
Give agents real power. Then a short leash.
Every action previews first, names any data loss, and stays inside its environment.
> preview plan: update web, run migrate · data loss: none
> deploy ✓ web.review-42.harbor.sh · 14s
> promote → production · keeps 12 secrets, 2 volumes
awaiting captain approval…- Preview before apply
- Data loss is named
- Scope is one environment
- You hold the promote button
BUILT FOR HOW TEAMS ACTUALLY SHIP
Six Tuesdays. Zero drama.
The deploy went sideways at 2am.
Roll back to the last snapshot. Go back to bed.
New checkout service.
Drop it on the canvas. Rope it to the production database.
The agent finished.
It previewed. You read the plan. You press promote.
Someone asks what changed.
Logs, metrics and history sit next to the deploy.
The bill arrives.
It's the server you already own. That's the bill.
WHY OWN THE SHIP
The cloud was a rental.
Rent is due.
A modern server is absurdly fast and absurdly cheap. Bring the €40 box, the rack, or the Ryzen you built for "machine learning".
- 1machine to start
- 0registries or YAML required
- $0for the runtime, forever
- 100%of your data on your disks
CHOOSE YOUR VESSEL
Rented yacht. Aircraft carrier. Or a ship.
Ployz is the boat in the middle: fast to sail, yours to keep.
| Hosted PaaS | Kubernetes | Ployz | |
|---|---|---|---|
| Runs on | Their regions | Your YAML | Your servers |
| Environments | App only | Build it yourself | Cloned, data included |
| Rollback | Redeploy and hope | helm rollback, maybe | Any snapshot, seconds |
| Deploys | A spinner | kubectl and prayer | A plan you read first |
| Exit | Rewrite | Rewrite | It's Docker. Walk away. |
THE TREASURY
Free to sail. Cheap to crew.
The runtime is free forever. Cloud costs less than a round of drinks.
Runtime, CLI, mesh, ingress. Yours.
Dashboard, unlimited servers, 2 environments.
Logs, backups, scheduled jobs, more environments.
Unlimited everything, priority support.
PARLEY
Fair questions.
Is this Kubernetes?
No. It's Docker machines, a deploy planner, a private mesh and a dashboard. If it runs a container, it runs Ployz.
What servers can I use?
Any Linux box you can SSH into. Start on one, enrol more with one command.
How far along is it?
Open beta. Deploy, preview and apply, clones, snapshots and rollback ship today. Review environments on every pull request are next.
THE INTERNET NEEDS MORE LITTLE ADVENTURES.