Everything you’d expect from an app host — one command to deploy, HTTPS on every app, ship on git push. The difference is the invoice: one flat monthly price, however many apps you run.
Cancel anytime · no lock-in · runs in the EU
Per-app hosting adds a line to your bill for every app you ship. Ours has one line.
| What you're running | Per-app hosting | Allocus |
|---|---|---|
| 1 small app | ≈ €7/mo | €10/mo |
| 5 small apps | ≈ €35/mo | €10/mo |
| 10 small apps | ≈ €70/mo | €10/mo |
Break even at two apps — after that it isn't close. At exactly one app we're the more expensive option, which is the honest version. The per-app column is roughly the cheapest always-on paid tier on the per-app platforms (about €5–7 per app, per month, databases extra); their current pricing pages have the exact numbers. Ours is one line on one invoice, however many apps you deploy.
Three commands and a sentence. No Dockerfiles to hand-write, no secrets to juggle, no DNS to wait on.
curl -fsSL https://allocus.dev/install | shallocus loginPaste the API key from your dashboard. Your box is already warming up.
claude plugin marketplace add allocusdev/allocusclaude plugin install allocus@allocusdeploy this to allocusLive at app.you.allocus.dev with a valid certificate, the instant it deploys.
A real session, start to finish: one sentence to Claude, and a
todo app is live on HTTPS. Sped up — nothing else cut.
Not a Claude user? Skip step 3 and run allocus init then
allocus deploy instead — same box, same result.
Read the full quickstart for every command.
The distinctive part: you don’t write the deployment. You ask for it.
The Claude Code plugin reads your repo and produces the Dockerfile, the
allocus.yaml, and the CI workflow — then runs the deploy. Multi-service
stacks (frontend + backend + database) work the same way.
Coming off Heroku, Scalingo or Render? Point Claude at the repo and say the sentence. A bare Flask, Django, Rails or Node app with no Dockerfile at all is fine too.
Claude is optional. The allocus CLI does everything on its own,
and a GitHub Actions workflow deploys on every push to main. Same box, same result.
Predictable by design: you pay for the box, not for each idea on it. Pick how much room you want underneath — every size runs unlimited apps.
Show off your ideas — five to ten small ones at once.
Find your first real users — one app with a database, plus the side projects.
Turn it into a business — several busy apps, or one heavy one.
Move up or down the ladder in a click — the disk is the same on every size, so a resize is a restart rather than a rebuild: your apps and your data stay put, and you’re offline for about a minute. Overload what you picked and it just slows down, never costs more; the dashboard shows the load, so you know when to size up. Cancel anytime from your dashboard — there’s nothing to unpick and nothing to email us about.
Including the ones where the answer is “not yet”.
A real virtual machine, provisioned for you and nobody else. It runs Docker and a Traefik edge that terminates TLS. Your apps are containers on your machine — not processes on a shared host, and not dynos that sleep when nobody’s looking.
Scaleway’s Paris region (fr-par). Scaleway is a French provider, so your
box and its data sit on EU infrastructure and stay there.
Your box keeps running until the end of the term you’ve already paid for — cancelling doesn’t cut you off mid-month. When that term ends, the box is destroyed along with everything on its disk. Export anything you want to keep before then. If you resubscribe later you get a fresh box, not the old one back.
Not yet — this is the honest answer. There are no automatic snapshots today, so treat the box the way you'd treat any single server you run yourself: keep your code in git, and dump anything precious somewhere else on a schedule. Nightly snapshots are on the roadmap and this answer will change.
Yes. Point the domain at your box — a CNAME for shop.myapp.com, an A record
for a bare myapp.com — list it under domains: in your
allocus.yaml, and deploy. The certificate is issued during that deploy and is usually
live within seconds; there's nothing to configure and no extra charge. Your
app.you.allocus.dev hostname keeps working alongside it.
How to set it up.
Yes, as part of your stack: run Postgres (or Redis, or whatever you like) as a container alongside your app, with its data on the box's disk. That's a genuinely normal way to run a side project. What doesn't exist yet is a managed database that lives independently of the box — so today, "the database" and "the box" share a fate. See the cancellation and backup answers above.
No. It's the fastest path and it's what we built the plugin for, but the
allocus CLI is a complete interface on its own, and the GitHub Actions workflow deploys
on push. If you never open Claude, everything still works.
Change tier from the dashboard. Upgrades are charged pro-rata straight away; downgrades credit you back. Either way the box restarts at the new size in about a minute and your apps and disk come back with it. Nothing to migrate.
Exactly that: there is no app count anywhere in the billing. The limit is physics — the CPU, RAM and disk of the size you picked. Fill it up and things get slower. They don't get more expensive, and you don't get an invoice you weren't expecting.
Fair enough. The CLI and the Claude plugin are public — read the code before you trust us with a card.