Pricing
We're still building this page out. In the meantime: RunOS is completely free today, unlimited nodes, unlimited clusters, no credit card needed.
Start freeProposed model, not live pricing
A percentage of what you already spend.
RunOS is $0 today: unlimited nodes, unlimited clusters, no credit card. This page is one proposal for what a future paid model could look like, worked out with real numbers so it can be judged on its actual shape, not a placeholder.
The rate card
One rate. No tiers, no seats, no per-service metering.
RunOS fee
10%
of your monthly infrastructure bill, whatever it is: Hetzner, AWS, bare metal, anything.
Who bills what
Two invoices
Your provider bills you directly, full price. RunOS bills the 10% separately. No markup baked into the infra bill.
Markup on hardware
$0
RunOS never touches your provider's price. The 10% is a separate line, calculated off it.
Worked examples
Same three cluster sizes we use everywhere else on this site, all on Hetzner, all compared to the equivalent AWS instances.
| Cluster | Hetzner bill | RunOS fee (10%) | Total | AWS equivalent | You save |
|---|---|---|---|---|---|
| 10 nodes 4 vCPU / 8GB each | $74/mo | $7.40/mo | $81.40/mo | $1,180/mo | 93.1% cheaper |
| 50 nodes 8 vCPU / 16GB each | $750/mo | $75/mo | $825/mo | $11,800/mo | 93.0% cheaper |
| 100 nodes 16 vCPU / 32GB each | $3,000/mo | $300/mo | $3,300/mo | $47,200/mo | 93.0% cheaper |
The savings percentage barely moves with scale, 93.1% at 10 nodes, 93.0% at 100. That's the point of a percentage-of-spend fee: it doesn't compress your margin as you grow.
Grounded in
None of the platforms we track (Porter, Northflank, Qovery, Humanitec, Railway, Render, Heroku, Vercel) price this way. Every one of them charges flat, per-seat, or metered-by-their-own-hosted-resource fees, because every one of them also hosts your workload. RunOS doesn't: it's BYOC, so the only number left to price against is your own infra bill. As far as we've found, no competitor prices as a straight percentage of the customer's own cloud spend, so this is the most novel model in this set, and the only one whose relative savings barely move regardless of cluster size or provider choice.