Home / Home / Integrations / Spend Control for Vapi

Spend Control for Vapi

Vapi builds AI voice agents that make and take calls — billed per minute of telephony plus LLM tokens per turn. At call volume, both lines compound fast.

Why Vapi agents overspend

Per-minute telephony charges scale directly with call volume — and with failed-call retries.

LLM tokens per turn multiply across long or looping conversations.

No built-in dollar budget for the agent's call spend.

How it works

Call sipi.bot before a call or a high-cost turn: amount, merchant, category → APPROVED, BLOCKED, or FLAGGED in ~5 ms, fully logged.

Rules that fit voice workloads

Per-call or per-day telephony ceiling.

Category rule: telephony vs inference.

Velocity limit so retry loops die before the minutes pile up.

Guard call

import requests

r = requests.post("https://sipi.bot/v1/transactions/evaluate",
    json={"amount": 12.5, "merchant": "vapi", "category": "telephony"},
    headers={"Authorization": "Bearer KEY"})
decision = r.json()["decision"]  # APPROVED | BLOCKED | FLAGGED

One call before the minutes add up.

FAQ

Does this slow call setup?

No — ~5 ms, and only spend actions trigger it.

Can I cap per campaign?

Yes — per-agent rules per campaign or use case.

Stop the next $12,400 night.

One API call (or MCP tool) in front of every agent transaction — APPROVED, BLOCKED, or FLAGGED, deterministic, ~5 ms, fully logged.

See plans — from $99/mo Try a live check