Cloud costs

Cut your cloud bill without breaking production.

We read the last invoices, name the drivers, and model options before anything moves. The numbers stay in the private report.

Get a free technical assessment

Cost lever check

Where does it run?
What does the workload look like?
Who looks after production?

Open the full cost lever check

Where the money goes

Function invocations and duration

Which routes run longest, and how often they wake.

Bandwidth and egress

What leaves the region, and whether a CDN would change that.

Image optimization

Originals served on every view versus sized variants.

Database compute tier

Idle capacity, and a staging database that matches production.

Preview and staging left running

Branches and servers nobody has opened in weeks.

Log retention and add-ons

Lines you pay to keep and tools you no longer open.

Evidence, not a promise

  • Flexera's 2025 State of the Cloud estimates 27% of IaaS and PaaS spend is wasted. Flexera 2025
  • 37signals has written about leaving a large cloud. That is a large-company example, not typical for a small team. 37signals on leaving the cloud
  • An extreme case: a traffic spike produced a Netlify $104K bill, later waived. Cybernews

Options

Same vendors. Spend caps, caching, smaller previews, and a database tier that matches the workload.

  • What you gain: Least change. You keep the current deploy.
  • What you take on: The unit prices of those vendors stay.
  • Who it suits: Teams who are happy with the host and want the waste removed.
Options only. Numbers from your invoices belong in the private report, not on this page.

Cloud cost cut audit

  • What it's for: teams whose Vercel, Supabase, AWS, Netlify, Firebase, Render or Heroku bill grew faster than users.
  • What you get: a line-item review of the last 3 invoices, a usage-driver map, a right-sizing list, caching and egress notes, spend caps and alerts, and migration options. Modeled cost stays in the private report.
  • Typical timeline: 5 business days.
  • What we need: a read-only billing export and a diagram of what is running.
  • What happens next: a proposal only if a change is worth making. Then a migration sprint if you accept it, and a care plan if you want the watch to continue.
  1. Free assessment
  2. Proposal
  3. Acceptance
  4. 30% advance
  5. Build
  6. 40% on completion
  7. UAT (your acceptance testing)
  8. 30% after UAT
Every price is a custom quote after your assessment. No work starts until you accept the proposal.

Custom quote after a free assessment. Terms: 30% to start, 40% on completion, 30% after your acceptance testing.

FAQ

Questions worth answering first

Will we have downtime?

In-place changes are planned to avoid it. A move uses a cutover window written in the proposal.

Is self-hosting safe?

It is safe when patching, backups and a restore test are someone's job. If that job does not exist, we say so.

What if moving costs more than it saves?

Then we say so, and the recommendation is to stay and right-size.

Do you need billing access?

A read-only billing export is enough for the audit.

Which providers?

Vercel, Netlify, Supabase, Firebase, AWS, GCP, Azure, Render, Railway, Heroku and Replit Deployments.

Next step

Not sure what your app needs?

Book a free 30-minute technical assessment. You'll leave with a short written list of risks and next steps, whether or not you hire us.

Every price is a custom quote after your assessment. No work starts until you accept the proposal.