method

The open source is free. The system is the product.

We build on a proven open source ecosystem — public Terraform modules, reusable components, the Atmos framework, all Apache 2.0, used in production by hundreds of teams. Standing on public, proven foundations is the opposite of the bespoke contraption only its author knows how to maintain. What you pay for is the architecture that holds the pieces together, implemented in your accounts.

And part of those foundations we maintain ourselves: what we publish is readable before you even talk to us.

Everything is yours, from day one

Your repositories, your accounts, your pipelines. No proprietary platform, no installed agents, no license to renew.

// if one day you send us away, you lose nothing.

Fixed price, guaranteed outcome

Scope and timeline defined before we start, with a refund clause written into the contract. We don’t sell days: we sell a result.

// whoever sells days has an interest in stretching. whoever sells an outcome, in finishing.

Infrastructure that runs, not slides

The deliverable is a platform in production with its documentation, not a roadmap in a PDF.

// pdfs don’t serve traffic.

We actually leave

The engagement ends with your team operating the platform: handover and training are in the price. Afterwards, if needed, a transparent support subscription.

// the measure of success is no longer being necessary to you.

speed

In production in weeks, not months.

Not because we rush: because we don’t start from zero. The stack is battle-tested (modules and architectures that ran in production long before reaching you) and we use AI to adapt it to your context: accounts, environments, constraints, conventions, policies.

The repetitive work is done by the machine; the decisions are made by someone accountable for them. Every line goes through human review and is written to be read three years from now. AI speeds up the craft, it doesn’t replace it: no slop in production.

🧱 battle-tested stack
🤖 ai for the tailoring
👀 human review on everything

deliverable

What’s inside, concretely.

The stack is well known and chosen to be operable by a small team. No exotic pieces to babysit.

landing zone: clear boundaries: multi-account on AWS, multi-project on GCP and Scaleway, multi-subscription on Azure
terraform + atmos: all infrastructure as code, organized to be read
kubernetes + argocd: EKS, GKE or AKS, declarative deploys: environments stay aligned with Git
karpenter: the right nodes, the right size, only when needed
promotions with kargo: from dev to production with written rules; the CI stays whatever you already use
observability: metrics, logs and alarms that only ring when they should
iam and sso: least privilege by default, access centralized and revocable
costs under control: allocation by team and service, budgets and alerts
documentation and runbooks: written for your team, not for us

if you operate in a regulated sector

GDPR, NIS2, DORA for finance, ACN and AgID if you work with the Italian public sector: a well-built platform doesn’t certify you, but it makes the evidence a by-product. Native audit trail, least privilege by default, and infrastructure as code is itself the documentation the auditor asks for.

the path

Three steps, with the roles in writing.

1.
Understand
you
you give us read-only access to billing and accounts, and half a day of your team’s time.
us
we come back with a prioritized report: what to recover, what to fix, in what order.
2.
Build
you
one technical point of contact, and the decisions that are yours to make.
us
we implement the platform in your accounts, via pull requests in your repositories, with scope and price fixed beforehand.
3.
Hand over
you
your team takes charge of the platform, with us alongside.
us
training, runbooks, shadowing. Then we step off stage; if needed, we stay reachable on a subscription.

who it’s for

Who it’s for, who it’s not.

we work well with

product teams between 15 and 200 people, on AWS, GCP or Azure
people who prefer Terraform to clicking around a console
people who want the platform to stay with their own team
anyone whose cloud bill has become a serious line item

we’re not the right vendor

if you’re looking for billable days to resell inside a bigger project
if infrastructure must forever remain an external department: we work to make ourselves unnecessary
if you need a mainframe migration or an ERP: not our trade

Saying no to the wrong sale costs less than managing it. That goes for you and for us.