See what a call costs.

Every call itemizes carrier, transcription, model, and voice, with the unit, source, and price-book version on each row.

Provider stacksper minute
Costat least $0.013/min

Twilio · price book 2026-08-05

Human-like and latency aren’t measured per stack yet. Cost is the price-book floor.

OpenAI Realtime (speech-to-speech)
  • Carrierconnects the call · Twilio, US inbound$0.0129/min
  • Realtime modelhears, reasons, and speaks in one hop · OpenAI gpt-realtime$32.00/1M input tokens (audio)$64.00/1M output tokens (audio)
Floorat least $0.013/minThe carrier row only. The realtime model’s audio tokens are excluded.
Streaming STT + LLM + TTS pipeline
  • Carrierconnects the call · Twilio, US inbound$0.0129/min
  • Transcriberhears the caller · Deepgram nova-3, streaming$0.0048/min
  • Modeldecides what to say · OpenAI gpt-4o-mini$0.15/1M input tokens$0.60/1M output tokens
  • Voicespeaks · Cartesia sonic-3$50.00/1M characters
Floorat least $0.018/minCarrier and transcriber only. Per-token and per-character rates are excluded.
Cheapest defensible pipeline
  • Carrierconnects the call · Telnyx, US inbound$0.0035/min
  • Transcriberhears the caller · Soniox stt-rt-v4, real-time$0.0019998/min
  • Modeldecides what to say · OpenAI gpt-4o-mini$0.15/1M input tokens$0.60/1M output tokens
  • Voicespeaks · Fish Audio s2.1-pro$15.00/1M UTF-8 bytes of input text
Floorat least $0.005/minCarrier and transcriber only. Per-token and per-character rates are excluded.

Price book version: 2026-08-05.

One call, itemized.

The product’s cost meter for the default stack — not a plan or a page estimate.

Default stack, per minute

From the product’s cost meter
What ranUnitSourceVersionAmount
Transcriberhears the callerper minuteSoniox2026-08-05$0.002
Voicespeaksper minuteCartesia2026-08-05$0.038
Modeldecides what to say · assumes ~1,403 in + 214 out tokens per minute, measuredper minuteOpenAI2026-08-05$0.009
Carrierconnects the callper minuteTwilio2026-08-05$0.0129

About $0.062 a minute on the default stack, at the providers’ published rates, itemized by layer on every call: carrier, transcription, model, voice. Run your own numbers in the cost calculator.

Costs live next to the agent.

The agent editor shows each component per minute and what is not included.

Agent editor · per minute

Transcriber · $0.002/min

Soniox — hears the caller.

Voice · $0.038/min

Cartesia — speaks.

Model · billed per audio-token

OpenAI — decides what to say. Not a rounded guess.

Carrier · at the carrier’s published rate

Twilio — connects the call.

Illustrative agent-editor cost strip in the builder’s format.

Straight answers about cost.

Callable

It depends on the stack. The default stack runs about $0.062 a minute at the providers’ published rates, itemized by layer on every call.

Callable

Each provider’s published price list, kept in a versioned price book with the vendor page linked on every row. Rate changes get a new version, not a silent edit.

Callable

Not yet. Today you pay for usage, itemized by layer at the rates on this page. If that changes, it changes with notice and a new price-book version.

Callable

Yes. Every call records the price-book version, so any charge can be checked against its rate.

Callable

A dash means the number isn’t set yet. We show that plainly instead of guessing or showing a zero.

Callable

No plan price is published. Usage is billed per call at the itemized rates above. New accounts start with $5 of free credit, no card required.

Callable

Connected minutes on transcriber, voice, model, and carrier are metered separately and itemized on the receipt.

Callable

Each call has its own dashboard receipt, itemized like the bill above, with the price-book version that billed it.

Callable

It depends on call volume. See the worked examples above for 100, 500, and 2,000 minutes a month, at the default-stack rate — not a monthly plan, since none is published yet.

What an AI receptionist costs per month on Callable.

Worked from the default-stack rate above, not a receptionist plan.

Worked examples, at the default-stack rate

100 minutes a month · about $6.20

100 × about $0.062/min, at the default-stack rate on this page (price book 2026-08-05) — at the providers’ published rates, itemized by layer.

500 minutes a month · about $31.00

500 × about $0.062/min, at the default-stack rate on this page (price book 2026-08-05) — at the providers’ published rates, itemized by layer.

2,000 minutes a month · about $124.00

2,000 × about $0.062/min, at the default-stack rate on this page (price book 2026-08-05) — at the providers’ published rates, itemized by layer.

Billed per call at these rates, not as a flat number or a plan.

See how Callable answers as an AI answering service. Read more: a buyer’s checklist for AI receptionist software, three ways to build an AI receptionist, and answering service pricing, compared plan by plan.

Describe the line. Read the bill.

Sign up and start with $5 of free credit. No card required.

Describe the line you need, then publish it today.

Create a Callable