We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
00 / pricing
Two plans. No seat math.
Cloud Pro when we run it, Business when you do. Members are unlimited on both. Cloud usage settles against a prepaid credit with a hard cap, and every rate that can appear on a bill is on this page.
Cloud Pro
For teams that want workflows running this week without operating a cluster. We run the control plane, isolated runners, upgrades and metering.
Includes a $50 monthly usage credit. $2,990/yr gets two months free; localized totals come from Paddle.
Secure Paddle checkout. Cancel anytime.
- Unlimited members, no per-seat pricing
- 10 concurrent runs and 25 GiB storage included
- gVisor-sandboxed runners we operate, patch and upgrade
- Managed AI at provider catalog cost + 10%; bring your own key at no extra charge
- Prepaid credit with a hard cap, on by default
Business
For teams whose data, credentials and execution must stay on their own infrastructure. The full platform on your Kubernetes or Docker.
Billed annually by contract. No usage meter: it runs on your hardware.
A scoping call with the founder, then a written license.
- Private self-hosted distribution and commercial license
- 1 production installation plus 1 non-production staging installation
- Unlimited members; workflow and app execution on your infrastructure
- Customer-owned AI and provider keys with no Rollout usage charge
- Product updates and direct support during the license term
01 / usage
The meter
Cloud Pro usage · settled in dollars
- runner time
- from $0.01/min · M
- S — 0.5 vCPU · 2 GB $0.005/min
- M — 1 vCPU · 4 GB $0.01/min
- L — 2 vCPU · 8 GB $0.02/min
- XL — 4 vCPU · 16 GB $0.04/min
- platform operations
- $0.20/1,000
- storage
- $0.20/GiB·month
- managed AI
- cost +10%
- your own keys
- $0
- included credit
- $50/month
Shell, container and code steps on a sandboxed runner, metered by the second.
One per step that runs without a runner: an API call, a record write, a branch. Query and search steps count one per 100 records returned.
The first 25 GiB are included every month.
The model provider's catalog price for the tokens a step used, plus an operations margin.
Model calls on your own provider key carry no Rollout usage charge.
Every rate above draws from it first; top up with prepaid credit.
Usage settles against prepaid credit in plain dollars, never points. The hard cap is on by default: once the credit is spent, Rollout stops admitting new runs and refuses model calls that cannot reserve credit, until you top up or the next month's credit arrives. Running out of credit deletes nothing.
02 / a worked month
What a month costs, line by line
| workload | metered as | cost |
|---|---|---|
| Cluster check every 15 minutes, 5 native steps (2,880 runs) | 14,400 operations | $2.88 |
| 100 builds of 8 minutes on an M runner | 800 runner minutes | $8.00 |
| The builds' 3 native steps each | 300 operations | $0.06 |
| Total, drawn from the $50.00 included credit | $10.94 | |
Arithmetic from the published rates, not a quote. $39.06 of the included credit is left for other work. Managed AI depends on the model and the tokens, so it is not in this example; with your own key it adds $0.
03 / compare
Which one is yours?
| Cloud Pro | Business | |
|---|---|---|
| Who operates it | We do: control plane, runners, upgrades | Your team, with our updates and support |
| Where steps run | Isolated cloud runners, plus your SSH hosts and registered desktops | Your Kubernetes or Docker, plus your SSH hosts and registered desktops |
| Members | Unlimited | Unlimited |
| Usage | Metered from prepaid credit, hard-capped | Unmetered |
| AI models | Managed catalog or your own keys | Your keys |
| Billing | $299/mo, cancel anytime | $24,000/yr by annual contract |
optional Business service
Design Partner implementation
Founder-led implementation of your first production process on Business: one scoped process, agreed success criteria, weekly working sessions and a clean handoff. It is consulting on top of the Business license, and only a few run at a time.
$4,000/mo
3-month minimum, scoped separately
Cloud add-on
Dedicated capacity
Runner nodes reserved for your workspace alone, in our cluster, with the same sandboxing and metering as shared runners.
- D4 · 4 vCPU / 16 GiB machine, about 3.9 vCPU / 13 GiB usable after system overhead · $499/mo or $4,990/yr
- D8 · 8 vCPU / 32 GiB machine · $899/mo or $8,990/yr
A single node is a single machine. Maintenance or a node failure pauses dedicated workloads until it recovers; add a second node for continuity.
Cloud Pro subscribers add dedicated capacity from workspace billing settings.
04 / fair questions
- How does Cloud usage billing work?
- Cloud Pro includes a monthly usage credit. Runner time, platform operations, storage beyond the included allowance and managed AI draw from it, then from prepaid credit you add. Your hard cap stops new work instead of creating an open-ended bill. Calls on your own AI keys carry no Rollout usage charge.
- What happens when the credit runs out?
- Rollout stops admitting new runs and refuses model calls that cannot reserve credit until you top up or the next month's credit arrives. Nothing is deleted. The cap is the ceiling on your spend, not a soft warning.
- Is Design Partner required for Business?
- No. Business is the product and license. Design Partner is optional consulting for a team that wants founder-led implementation of its first production process.
- Can we buy Business month to month?
- No. Business is an annual self-hosted license because private deployment, updates and support need a longer operating relationship. Design Partner is billed monthly with the minimum term shown above.
- Do you offer Personal, Team, or Enterprise?
- Not today. Cloud Pro covers managed workspaces without seat pricing and Business covers licensed self-hosting. Unusual security or support requirements are quoted against Business rather than an undefined Enterprise tier.
More on models, security, self-hosting and MCP is in the FAQ .
Still deciding?
Twenty minutes with the founder answers more than a pricing table. You talk to the person who built it.
Talk to the founder