Your AI assistant spends more on its brain than its voice.
real · by action
Brain · · memory sync + LLM + analysisVoice · · Vapi minutes + speech in/outfrom the Vapi end-of-call breakdown
Where the money goesby model
all-time
One model carries the bill.
Choose a view
Saved for this lens.
Cost vs the revenue it rides onreal
of monthly revenue
Monthly revenue in scope
·
of monthly revenue
This month's AI cost against the monthly revenue in scope.
Choose a view
Saved for this lens.
Cost to serve, per client real · from calls
Per-call cost attributed to each client from the end-of-call breakdown. Unattributed means a call on a number not yet mapped in call routing.
Provider budgets and alarms Live
Postpaid providers alarm on month-to-date spend and error rate; a prepaid provider alarms on remaining credit. Thresholds are set per provider. Never a kill switch: a live call is never cut.
Cost by the job the AI is doing real
Untagged is the biggest bucket until each event carries the job it ran for.
Budget spend real
used this month
·cap $50
The cap and its alert thresholds are set in Settings. Alerts at 80% and 100%. Never a kill switch: a live call is never cut.
Every cost event real
When
What
Ran on
Model
Persona
Cost
Reading.
Cost to serve, per call real
When
Client
Channel
Top model
Parts
Cost
Reading.
Platform overhead actual · monthly
infraFixed monthly platform costs the per-call cost rides on. Dev tooling is shown separately and is not cost-to-serve.