Every bill.
One statement.

AWS, Azure, Google Cloud and Oracle Cloud, read from each provider’s own cost export rather than a polled API — so you get full resource-level detail and a year of history, not a summary.

OpenAI, Anthropic and GitHub, so model and CI costs sit in the same statement as your infrastructure instead of arriving separately on a credit card.

Microsoft 365 and Okta bring seat counts in, which is what makes unused-licence waste visible at all.

Every source is mapped onto FOCUS 1.4, the open billing standard — which is why a Google Cloud line and an Oracle line can sit in the same table and actually be comparable.

Total spend, trend, and the biggest services across every connected source at once — kept per currency, never summed across them, because a CAD total and a USD total are not the same number.

Click a service and see which resources actually drove it, down to individual line items. A cost you can’t trace is a cost you can’t argue with.

Slice spend by account, by tag, or by your own custom dimension — with untagged spend shown as its own honest bucket rather than quietly redistributed.

Define groupings the cloud doesn’t know about — team, product, environment, customer — using ordered match rules. Cross any two of them to get a pivot table (team × environment, say). Shared costs can be split across members by percentage or in proportion to usage.

One toggle switches every figure between what you were billed and the amortized view that spreads commitments across the period they cover.

A projection of where the month lands, based on the spend pattern so far rather than a flat average.

Spikes and newly-appearing services surface automatically, measured against each source’s own trailing baseline. Anomalies can be detected per dimension member too — so a spike inside one team shows up even when the org total looks calm. Dismiss one and it stays dismissed for that day only, so a genuinely new spike next week still reaches you.

Set a budget or a spike threshold and get it in Slack. Scope a rule to everything, one cloud, one tag, or a single dimension member — and the alert names the largest driver, not just the number.

Upload a business metric — customers, orders, active users — as a two-column CSV, and see cost per unit over any period. Spend going up is only bad news if it isn’t buying more of something.

Track the SaaS that never appears on a cloud bill, with renewal dates and seat counts, so the annual invoice stops being a surprise.

Rightsizing, idle resources and commitment advice from all four clouds in one place, rather than four consoles nobody logs into.

Unused licence seats, renewal windows worth renegotiating, AI batch-processing opportunities, and non-production resources billing through the weekend. Each is labelled as computed by Capitops, so you always know whose recommendation you’re reading.

Every recommendation carries a rough effort rating, because the cheapest saving isn’t always the one worth doing first. Dismiss what doesn’t apply and it stays gone.

How much of your usage is covered by commitments, and your effective savings rate — the blended discount you’re actually achieving versus list price.

Recommendations show which dimension owns the affected resource, so the work lands with the right team.

“What did we spend last month, and how does that compare to the month before?” The agent reads the same data every other page shows, answers with real figures, and links to the view that proves each one. It’s read-only — it cannot change anything in your accounts.

Describe the table you want once; it’s saved under a name and refreshed daily, so the question you ask every Monday answers itself.

A plain-English summary of the month’s spend, what moved, and what to look at — delivered to Slack.

Export any period to FOCUS-standard CSV, or connect an AI tool of your own to the read-only MCP server and query your spend from wherever you already work.