> For the complete documentation index, see [llms.txt](https://docs.chatlayer.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.chatlayer.ai/voice/phone-and-voice/voicebot-speech-to-text-fine-tuning.md).

# Voicebot speech-to-text

## Configure speech-to-text for your voicebot

When callers speak to your voicebot, Chatlayer converts speech to text with a speech-to-text engine.

You choose the provider and model per **Sinch Voice** channel from the channel **Settings** tab.

### Open Sinch Voice channel settings

Speech-to-text settings apply to a **Sinch Voice** channel only.

Other channel types use their own settings elsewhere.

{% stepper %}
{% step %}
Go to **Channels**.

Open the **Sinch Voice** channel you want to configure.
{% endstep %}

{% step %}
Open the **Settings** tab.
{% endstep %}
{% endstepper %}

### Choose an STT provider

Under **Speech-to-text**, you can configure:

* **STT provider** — the engine that transcribes caller speech.
* **Google STT model** — shown only when **Google** is selected.

Use **STT provider** to pick the engine for this channel:

* **Microsoft** — default option. Uses Microsoft speech-to-text.
* **Google** — uses Google Cloud Speech-to-Text v2. You can also choose a Google model.

{% hint style="success" %}
Changes save immediately when you select a new provider.
{% endhint %}

{% hint style="info" %}
STT provider is set per **Sinch Voice** channel.

If you run multiple voice channels, each channel can use a different provider.
{% endhint %}

### Choose a Google STT model

When you select **Google** as the provider, an extra dropdown appears.

Choose the model that fits your use case:

* **Chirp 3** — default Google model. Good for most voicebots.
* **Chirp 2** — earlier Chirp model. Useful for specific compatibility needs.
* **Telephony** — optimized for phone-call audio.

If you switch back to **Microsoft**, Chatlayer does not use the Google model until you select **Google** again.

If speech-to-text fails during a call, review the issue in [Execution logs](/navigation/history/execution-logs.md).

### Improve recognition with phrase lists

You can submit a custom phrase list to improve recognition accuracy. To create a phrase list, open a support ticket and submit your list there.

This works well for names, addresses, product names, and domain-specific terms. For example, if your list includes `Chatlayer`, recognition prefers `Chatlayer` over `chat layer`. <br>

You can submit phrase lists:

* per bot language
* or globally across all bot languages

Each list supports up to **500 phrases**. Submit the list as a plain text file. Separate each phrase with a semicolon: `;` Remove duplicate phrases before sending.

### Related settings

On the same **Sinch Voice** channel **Settings** tab, the **Recording** section lets you enable **Save recordings** to store full call recordings automatically.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.chatlayer.ai/voice/phone-and-voice/voicebot-speech-to-text-fine-tuning.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.