See your AI spend before it gets out of hand.
Set a token or dollar budget per project or agent. Usage aggregates in real time, with a warning at 80% and a critical alert at 100% — fired once, not on every request.
Screens
What it does
- 01Per-project budgets
- A Source × Scope key scopes a budget to one project or agent. Leave both blank for an everything budget.
- 02Real-time monitoring
- Every new usage event triggers a fresh aggregation, so the dashboard is always current.
- 03Alerts fire once per period
- Warning at 80%, critical at 100% — only the moment a budget newly crosses the line, resetting each period.
- 04Automatic cost from pricing
- Ships with default per-million rates for major models. Edit them to match your actual contract.
- 05Weekly / monthly reports
- Change the date range in Usage and the same summary works for either cadence.
- 06Usable from any AI system
- comet-taskAI, ai-scheduler, your own scripts — anything that can POST over HTTP can log usage.
How it works
Check pricing, create a budget
Review model rates in Pricing, then set a per-project budget in Budgets.
Configure your AI system
Issue an API key and have it POST usage after every run.
Watch the dashboard
Progress bars tell the story at a glance; alerts arrive when a threshold is crossed.
Good to know
- One alert per threshold per period — no repeat once a level has fired.
- Alert state resets automatically on a daily / weekly / monthly boundary.
- Usage is always recomputed from source data — no counter drift.