You should not be paying us a margin on AI.
We charge for the product. The work runs on hardware you already have, and your AI spending goes straight to your own provider, which keeps our incentives pointed the right way.
Three lines, and two of them are not ours.
The product itself: approvals, limits, reporting, integrations and the record of everything. Priced by how many teams use it and how many people run and approve work.
Free. The work happens on machines you already have, so there is nothing for us to mark up.
On your own accounts, at your own rates. We can bill it through if you would rather, but the default is that what you spend on AI stays between you and your provider.
A vendor whose revenue goes up when your AI bill goes up has a quiet reason not to make it cheaper. Ours does not, which is why making each job faster and cheaper is something we work on rather than something that would cost us money.
Priced per team, not per repository.
A team can have as many repositories as it needs. Charging per repository would penalise you for exactly the setup we are best at helping.
One team, one machine, one approval path. What a first deployment looks like.
- · One team
- · Up to 5 people running and approving work
- · Standard hosting
- · GitHub and email
- · The full record of everything it did
Several teams, more machines, and different approval rules for different kinds of work.
- · Several teams, with cost broken down per team
- · Unlimited people who can watch; you pay for those who act
- · As many machines as you need
- · Different limits per team and per repository
- · Single sign-on through your own provider
- · Your own AI accounts and keys
Everything in your own cloud account, with the compliance requirements that come with that.
- · Everything in Business
- · Your own dedicated instance
- · Data residency and retention controls
- · Encryption keys you manage yourself
- · Fits your existing change and audit process
- · Named support with an engineering path
No numbers on the page, and not out of coyness. What you pay depends on how many teams use it, how many people run and approve work, whether you want your own instance, and whatever your compliance requirements add. A table pretending otherwise is just a figure we would change on the call anyway.
What you can budget for: a fixed annual cost that does not move with how much you use it, plus your own AI spending, where you can already see and cap it.
- It matches the value
One repository with 400 engineers behind it is not worth less than 300 small ones. Repository count predicts nothing.
- It matches our cost
Each team needs setting up and supporting. That is where the real work is on our side.
- Nobody games it
You do not reorganise your teams to move a bill. Teams follow how you work, not how you are charged.
Tell us the shape of your estate.
How many teams, how big, and how approvals work today. That is enough for a firm number with the assumptions written next to it.