Your cloud and AI spend, reviewed before you wake up.

Every run teaches nable more about your environment, so it gets sharper about what is normal for you and what is not. You wake up to a read of what happened overnight across AWS, Azure, GCP and your model spend, already analysed, with a draft fix attached to the few things worth acting on.

20 providers · read-only access · nothing shipped without a human. Install with uvx nable.

Reads billing from

It finds the money, explains it, then ships the fix

Catch it: every series is watched against its own pattern, not a static threshold, so you hear about the spike the morning it happens with the cause attached. Explain it: plain English with the working shown, including what nable decided against and why. Fix it: a pull request against your infrastructure repo, scoped to one change, with the saving and the blast radius in the description.

Where the brief lands

One nightly run, four places to meet it: a Slack channel post with the drafts, the morning brief in your inbox, the same run over the API, or a CI check that fails a plan adding spend you did not expect.

A dashboard for the people who own the number

Finance gets budgets, commitments, chargeback and a savings ledger that only counts changes that actually shipped, auditable line by line.

What nable can and cannot do

Scoped to billing data: nable connects with a read-only role scoped to billing and usage data, and never touches application data, customer data or anything in your production databases. It drafts, you decide: write access is opt-in per provider and per action, and the default is to look, explain, draft a change and wait for a human. Savings you can audit: a change only counts once it shows up on an invoice.

Set up in about four minutes

One command, one provider, one question. Run uvx nable, then nable connect aws with a read-only role, then ask nable ask "what grew fastest last month?". nable backfills 18 months of history and has something to say by the next morning.