CostMon

What problem does it solve?

Spend is scattered across a dozen cloud, AI, and SaaS billing consoles. Each has its own format, so nobody has one trustworthy total.

Who is it for?

Engineering and finance teams tired of reconciling a dozen separate billing consoles.

Why are we building it?

We're a studio that runs nine products, plus every client's stack. Our own AI and cloud spend already lands across a dozen billing consoles, and so does every SaaS subscription. Each has a different export format and a different definition of "usage." Reconciling that by hand every month is exactly what we'd tell a client to stop doing.

CostMon is the fix we're building for ourselves first. It gives one normalised view of spend across providers. Engineering and finance then look at the same number, instead of two spreadsheets that never quite agree.

It's still in build. That's not because the idea is unproven; we already do this reconciliation by hand. It's because normalising a dozen billing formats into one honest total is worth taking the time to get right before anyone else depends on it.

How does it work?

1. Provider connectors, not manual exports

   Every provider we bill through becomes a connector. It pulls billing data on its own schedule, so the aggregation never depends on someone remembering to download a CSV.

2. A normalised spend schema

   Every provider's invoice gets mapped into one common shape: line item, service, cost, and period. That makes a dollar of AI inference spend and a dollar of cloud compute spend comparable.

3. One trusted total, not nine partial ones

   The normalised records roll up into a single view engineering and finance can both look at, instead of each side maintaining its own reconciliation.

What's the hard part?

Every provider bills differently: different line-item granularity, different currencies and proration rules, different definitions of what counts as a "unit." The hard part isn't pulling the data. It's designing a normalisation schema honest enough to make a cloud invoice and an AI usage invoice comparable. It must not quietly lose the detail that explains a cost spike.

What does it do?

• Aggregates billing from every cloud and AI provider
• Normalises spend into a single comparable view
• One trusted total across all providers
• Shared source of truth for engineering and finance

What does this prove we can do for you?

This is the same billing-normalisation and cost-observability work we do for clients asking what their AI spend actually looks like. CostMon is us solving it for ourselves and every other studio product first.

Where this shows up in client work

• What AI actually costs
• AI cost benchmark by workload