Know what every run costs — and cap it before it surprises you.
Every span carries a price. Cost rolls those prices up by feature, by model, and by day — then a hard monthly budget refuses any run that would push you past your limit. No end-of-month invoice you didn't see coming.
$842 / $1,000
near capby feature
01
Budgets with teeth, not just alerts
Set a soft limit for an early warning and a hard limit that actually stops spend. Once the month's total reaches the hard cap, the next run is refused with a clear error instead of quietly running up the bill. The same check guards the dashboard and the public API, so a runaway script can't out-spend your laptop.
$842 / $1,000
near capby feature
02
Spend you can attribute to the dollar
Cost is the sum of every trace and eval run, priced span by span. Break this month down by feature — Agent runner, Canvas, API, Evals — and by the exact model that did the work, then see your most expensive traces ranked with a link straight into the viewer. You always know where the dollars went.
03
A 30-day trend, zero-filled
A continuous 30-day sparkline shows this-month versus last-month spend and your average per active day, so a slow creep is obvious long before it becomes a number you have to explain. Quiet days read as zero, not gaps — the line never lies about a lull.
$842 / $1,000
near capby feature
04
Breaches notify the people who care
When a team crosses its hard cap, Cost emits a signed budget.breached webhook and a mobile push — wire it to Slack, Discord, PagerDuty, or your own infra. Every budget change is written to the compliance audit log, so the trail of who raised the limit and when is always there.
in email dana@acme.io, card 4242 4242 4242 4242
out email <REDACTED:email>, card <REDACTED:credit_card>
UI + public API
Enforced at
webhook + push
Breach signal
feature · model · day
Attribution
Part of one platform
Cost works hand in hand with Observability.
Speed plus trust — prove your agents got better this week.
Cost is one piece of Stackon, the observability-first workspace for teams running Claude and Codex. Start free and instrument your first run today.