Skip to content
Closed
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 28 additions & 1 deletion fern/pricing.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -70,7 +70,7 @@ For large-scale deployments, we offer:
</Accordion>

<Accordion title="How is billing calculated?">
Billing is calculated based on actual usage - voice minutes, API calls, and storage used. You can monitor usage in real-time in your dashboard.
Billing is calculated from usage across Vapi platform fees and provider passthrough costs (for example, transcriber, model, voice, and telephony). Some provider pricing is per stream, so dual-stream configurations can cost more than single-stream configurations. You can monitor usage in real-time in your dashboard.
</Accordion>

<Accordion title="Can I change plans anytime?">
Expand Down Expand Up @@ -112,6 +112,33 @@ For large-scale deployments, we offer:
</Card>
</CardGroup>

### Azure STT pricing clarification

Azure’s published real-time Speech-to-Text base price is typically around **$1.00/hour** (about **$0.0167/min**) **per stream**.

If you see **$0.016/min** in examples, treat it as a **single-stream reference rate**, not the default total for dual-stream calls.

By default, Vapi transcribes **two streams** for many voice pipelines:

- **Customer audio stream**
- **Assistant audio stream**

Because both streams are sent to Azure STT, the effective rate is typically about:

- **$0.033/min total STT** (2 × $0.0167/min), before any regional/provider-specific differences

<Note>
For streaming STT, Azure bills based on audio sent to the service. In continuous streaming, silence and pauses are still part of the streamed audio, so your billed minutes usually track stream duration.
</Note>

To reduce Azure STT cost toward single-stream pricing:

- Enable **`modelOutputInMessagesEnabled: true`** to skip assistant-channel retranscription when supported.
- This optimization currently requires **ElevenLabs voice** (word-level timestamps are needed for interruption handling).
- If you use other voice providers, the assistant stream may still be transcribed.

If you use **BYOK (your own Azure credentials)**, Vapi does not charge Azure STT passthrough on your behalf; Azure bills you directly in your own account.

### Starter Credits

Every new account is granted **$10 in free credits** to begin testing voice workflows. You can [begin using Vapi](/quickstart/phone) without a credit card.
Expand Down
Loading