Migrations
What moving off usage billing actually looks like
A real savings number, in writing, before you spend a dollar. You tell us what you're running and where it's hosted; we tell you what it'll cost to move and what you'll save. Then we do the move ourselves.
What does your bandwidth cost, per provider?
Published egress rates, compared on the same bytes. Not a capacity claim — a pricing comparison.
Log-scaled: 1,000 — 10,000,000 users/mo.
Outbound transfer: —
Assumption in use: ~15 MB/user/month. Ballpark planning figure, stated openly and never measured — adjust the profile to match your app.
Moves the Google Cloud figure only, because Google prices internet egress by destination continent and by nothing else — $0.02/GiB within North America against $0.14/GiB to South America, per cloud.google.com/vpc/network-pricing, checked 2026-09-12. AWS and Azure price by the region your server sits in instead; both are shown at their North America rates.
Both are our real published terms. Overage is cheaper when bandwidth is all you are short of; more instances is what you actually buy when compute is the constraint — and bandwidth alone cannot tell you which, which is why this is your choice and not our assumption.
Chart could not be loaded. The cost estimates below are unaffected.
Est. Olympia cost/mo
—
This compares published outbound-transfer pricing. It is not a guarantee of what any single server can handle — tell us your stack for a real quote. Competitor figures cover bandwidth only, plus a plan fee where the provider charges one; compute, storage, load balancers, support and everything else are excluded, so each is a floor rather than a bill. These vendors also do not agree on what a gigabyte is — AWS and Azure publish per GB, Google and DigitalOcean per GiB — and we quote each in its own unit rather than converting, which leaves about 7% of slack between them.
Sources
Who is not on this chart, and why
And what does your team cost, per seat?
Drag the team size. Every price is the vendor's own published rate, dated and linked.
1 — 500 people. Per-seat pricing is linear, so the gap widens with every hire.
One server, running the open-source replacements. The price does not move when you hire someone, which is the whole of the argument.
Per-seat tools, monthly
—
One Olympia server
—
Difference
—
The Olympia figure is the rent on a server. It is not the cost of running these applications. Installing them, upgrading them, patching them, taking backups, watching them and restoring them at 2am when something breaks is engineering time, and engineering time is money this block does not attempt to price. If nobody on your team wants that job, the honest comparison is the server plus a managed engagement — not the server on its own. Per-seat tools also buy you support, compliance paperwork and somebody else's on-call rota, and dropping them is a real trade, not a pure saving.
Sources
Want per-tool control, price overrides and a storage comparison? Use the full savings calculator.
Common migrations we handle
Every one of these ends the same way: you tell us the stack, we quote the savings in writing, then we do the move.
Running something that's not on this list? We still want to hear about it.
The short version
Here's what happens after you tell us your stack.
01
Tell us your stack
What you're running, where it's hosted, and what it costs you now.
02
We quote the savings
A real number, in writing, before you pay anything. No card required.
03
Order your server
Configure the tier that fits your quote and check out. It's provisioned within 24 hours of your order.
04
We do the move
App, data and DNS, on a window agreed with you. We tell you the timeline in writing before you commit to it.
Moving off something?
Tell us what you’re running now and we’ll scope the move and quote it in writing, before you pay for anything.
Not sure yet? Ask us a question.