> ## Documentation Index
> Fetch the complete documentation index at: https://www.bolna.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# What a Bolna voice AI call costs

> A one-screen summary of Bolna Voice AI pricing: the flat per-minute rate, what sits outside it, telephony charges, and where agent and concurrency costs apply.

A short answer for evaluators. The detailed breakdowns live in [Call pricing](/docs/pricing/call-pricing), [Agent pricing](/docs/pricing/agent-pricing) and [Preferred models](/docs/pricing/preferred-models).

## The short version

<CardGroup cols={3}>
  <Card title="Voice AI" icon="waveform-lines">
    Transcription + LLM + text-to-speech. Covered by a **flat per-minute rate** as long as your agent uses [preferred models](/docs/pricing/preferred-models).
  </Card>

  <Card title="Telephony" icon="phone">
    Charged per minute by the carrier, and it varies by country and provider. Billed by Bolna, or by your own provider if you [bring your own account](/docs/supported-telephony-providers).
  </Card>

  <Card title="Platform fee" icon="server">
    Bolna's own per-minute component. See [Call pricing](/docs/pricing/call-pricing) for how the three parts add up.
  </Card>
</CardGroup>

<Info>
  Billing is usage-based and prepaid from your wallet. Larger top-ups move you to a lower effective per-minute rate.
</Info>

***

## What pushes the cost up

| Choice | Effect |
| - | - |
| A model outside the preferred list | That component leaves the flat rate and is billed at variable provider rates — the most common reason two similar agents cost differently. See [Preferred models](/docs/pricing/preferred-models). |
| Long calls | Every component except the LLM is billed by duration. Tightening the prompt and setting a `call_terminate` limit in the [Call tab](/docs/agent-setup/call-tab) is the fastest lever. |
| Verbose LLM output | The LLM is billed by tokens generated. |
| Premium voices and international numbers | Both are priced above the standard bundle. |

For a working method rather than a table, see [Control cost per call](/docs/production/control-cost-per-call).

***

## What is not billed per minute

* **Agents themselves** — see [Agent pricing](/docs/pricing/agent-pricing) for how many agents a workspace can hold.
* **Concurrency** — how many calls you can run simultaneously is an account setting, not a per-call charge. See [Plan concurrency](/docs/enterprise/concurrency-management).

***

## Next steps

<CardGroup cols={3}>
  <Card title="Full pricing breakdown" icon="magnifying-glass-dollar" href="/docs/pricing/call-pricing">
    Every component, with worked examples
  </Card>

  <Card title="Preferred models" icon="list-check" href="/docs/pricing/preferred-models">
    What the flat rate includes
  </Card>

  <Card title="Control cost per call" icon="scissors" href="/docs/production/control-cost-per-call">
    Reduce spend without hurting quality
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.