# Voice Bot

### What is Voice Bot?

The **Voice Bot** in BotPenguin enables seamless, two-way **voice-based communication** between customers and bots. It allows users to interact over calls or smart devices, combining natural conversation with automation for a richer customer experience.

{% embed url="<https://www.youtube.com/watch?v=sfZ5QaZZqQ8>" %}

***

### Benefits of Voice Bot

* 🗣️ Enables **hands-free communication** through natural voice interactions.
* ⚡ Enhances **customer engagement** with real-time voice response.
* 🎯 Integrates advanced **AI and speech models** for accurate recognition.
* 🔊 Offers **noise cancellation** and **voice clarity** for professional-grade experience.
* 🧩 Provides **easy configuration** with multiple voice and transcription providers.

***

### How to Use Voice Bot

1. Log in to your **BotPenguin Dashboard**.
2. On the left-hand sidebar, click on **“Agents”** to open a previously created agent.\
   If no agent exists, refer to the documentation on [**Creating an Agent**](broken://pages/0MF1UDozkRvVprXxkdpz).

<figure><img src="/files/CNkhp2izaU6Z7uEP2TR5" alt=""><figcaption></figcaption></figure>

3. Once inside your selected agent, go to **“Agent Settings.”** Under this, navigate to the **“Voice”** tab (fifth sub-tab).

<figure><img src="/files/aBmzXRDzwSYHLSuzkLAi" alt=""><figcaption></figcaption></figure>

4. In the **General** section, enable the **“Voice Mode”** toggle. This activates the Voice Bot for your agent.
5. Configure the **General Settings:**
   1. **Enable VoiceMode:** Turns the Voice Bot ON.
   2. **Background Noise Cancellation:** Removes background noise for clearer communication.
   3. **Numerical Conversion:** Converts spoken numbers into digit format (e.g., “one” → “1”).
   4. **Voice Call Sound:** Choose the sound played during an incoming voice call to the agent.
   5. **Silence Timeout:** Defines how long both bot and user can remain silent before the call auto-disconnects (options: 10s, 30s, 60s).
   6. **Maximum Call Duration:** Sets total call length limit (options: 5s, 180s, 600s).

{% hint style="info" %}
Use **“Reset to Default”**, **“Update”**, or **“Update and Next”** as required.
{% endhint %}

<figure><img src="/files/XCrDexdXqOb4zEpHBOnO" alt=""><figcaption></figcaption></figure>

6. Move to the **Advanced Section.**
   1. Toggle **“Enable Advanced Voice Mode”** to allow users to interrupt the bot mid-conversation.
   2. Define **“Number of Words Before Agent Stops Speaking”**, determining how many user words will trigger the bot to pause and listen.
   3. Click **“Update”** to save changes.

<figure><img src="/files/6d4K2OkcmX9SjWo0BFb9" alt=""><figcaption></figcaption></figure>

6. Proceed to **Advanced AI Settings** to complete Voice Bot setup. Here, three integrations are mandatory:
   1. **LLM Provider** – The AI engine that powers conversational intelligence. The following should be integrated if the AI Agent has previously been made.
   2. **Voice Provider** – Handles the voice synthesis and interaction layer.
   3. **Transcription Provider** – Converts voice to text for recording and analysis.

<figure><img src="/files/ceGdf5aw8XVJImrIpeoe" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/lSofuTCjV25xxyfIJ68h" alt=""><figcaption></figcaption></figure>

***

**Voice Provider Setup:**

* Select your **Voice Provider**: *ChatGPT, 11Labs, Hume,* or *DeepGram.*

{% hint style="success" %}
🏆 *ElevenLabs and Hume are recommended for best results.*
{% endhint %}

* Choose **Voice Provider Credentials** (API key integrated via the [Integrations](broken://pages/8ZT8vpqoUzEn9BmEJTj2) section).
* Select **Voice** and **Model** based on provider:
  * DeepGram → Aura 2, Thalia EN, etc.
  * ElevenLabs → ElevenMultilingual V2, ElevenTurbo V2, Turbo V2.5, etc.
  * ChatGPT → GPT-40mini-TTS, TTS-1, TTS-1-HD.
* Click **“Update”** to save.

<figure><img src="/files/2paZ88xheRbGtGds4eqc" alt=""><figcaption></figcaption></figure>

***

**Transcription Provider Setup:**

* Choose **Provider:** *DeepGram*, *ElevenLabs*, or *ChatGPT.*

{% hint style="success" %}
🏆 *DeepGram recommended for best results.*
{% endhint %}

* Select **Transcription Provider Credentials** (API key from Integrations).
* Select **Model** and **Language** (supports English, Spanish, German, Turkish, Hindi, French, Russian, etc.).
  * Enable **Multilingual** if multiple languages are expected.
* Click **“Update”** to confirm setup.

<figure><img src="/files/9QIG3X4QW3QFWFKxpaih" alt=""><figcaption></figcaption></figure>

***

8. Configure Live Agent and End Call options (optional).

These configurations are made through the **Conversation Flow Instructions** located under **Agent Personality and Instructions → Prompt.**

<figure><img src="/files/ernJAI7PHUDWj7Vdf8Xi" alt=""><figcaption></figcaption></figure>

* In the **Conversation Flow Instructions**, use tools accessible by typing a backslash (`\`).
* Two tools are available for the Voice Bot:
  * **Transfer Call to Live Agent:**

    * In the instructions, type “\” and search for **Transfer Call to Live Agent**.

    <figure><img src="/files/05c99Bzu75zBP5x3TA9v" alt=""><figcaption></figcaption></figure>

    * Click on it to open the configuration prompt.
    * Enter the trigger phrase (for example, “connect me to a live agent” or simply “live agent”).
    * Click **“Add and Next”** to finalize setup.\
      When users say the configured phrase during a voice conversation, the bot will automatically initiate a transfer to a live agent.

    <figure><img src="/files/tZzJNkLUwJ0RtkUuPZp9" alt=""><figcaption></figcaption></figure>
  * **End Call Tool:**

    * Similarly, in the same **Conversation Flow Instructions** section, type “\” and search for **End Call**.

    <figure><img src="/files/VQjKvhEpAd8qmFyobiD9" alt=""><figcaption></figcaption></figure>

    * Click to open the configuration prompt.
    * Enter the words or phrases (e.g., “end call” or “goodbye”) that will act as triggers for ending the call.
    * Click **“Add and Next”** to finalize setup.\
      This ensures the bot terminates the call gracefully when the user intends to end the session.

    <figure><img src="/files/b6MUv1VJr4m7pOcyGvRK" alt=""><figcaption></figcaption></figure>

9. To start using your Voice Bot:

* Go to your agent’s main screen.
* Click the **three dots (⋮)** beside your agent.
* Select **“Call Agent”** to initiate the Voice Bot interaction.

<figure><img src="/files/8OzgJ8jkwJY02XGOWWXY" alt=""><figcaption></figcaption></figure>

During the call, two key options are available:

* **Mic:** Mute or unmute the call.
* **Cancel:** End the ongoing call session.

<figure><img src="/files/tByFOPv2F6HWu96pq0V0" alt=""><figcaption></figcaption></figure>

***

### FAQ's

<details>

<summary>What is the Voice Bot feature used for?</summary>

The Voice Bot enables your BotPenguin agent to communicate with users through real-time voice conversations instead of text chat.

</details>

<details>

<summary>Can I use Voice Bot without integrating a voice provider?</summary>

No. You must integrate at least one supported voice provider (such as 11Labs or DeepGram) for the Voice Bot to function.

</details>

<details>

<summary>What happens if the Silence Timeout limit is reached?</summary>

If both the bot and the user remain silent beyond the defined limit (10s, 30s, or 60s), the call automatically disconnects.

</details>

<details>

<summary>Can users interrupt the bot while it is speaking?</summary>

Yes. If Advanced Voice Mode is enabled, users can interrupt, and the bot will stop and start listening immediately based on your configured word count threshold.

</details>

<details>

<summary>How can a user talk to a live agent using the Voice Bot?</summary>

In the Conversation Flow Instructions, add the Transfer Call to Live Agent tool using the backslash (\`\\\`). Configure trigger phrases like “connect me to a live agent.” When the user says this, the call will automatically transfer to a live agent.

</details>

<details>

<summary>Can users end the Voice Bot call by voice command?</summary>

Yes. You can use the "End Call" tool in the Conversation Flow Instructions and define the user phrase (e.g., “end call”) to terminate the conversation automatically.

</details>

<details>

<summary>What providers are best for voice synthesis and transcription?</summary>

For voice synthesis, ElevenLabs and Hume deliver the most natural-sounding voices. For transcription, DeepGram provides the highest accuracy.

</details>

<details>

<summary>Can I use multilingual transcription?</summary>

Yes, you can enable multilingual transcription to support languages such as English, French, German, Hindi, Russian, and more.

</details>

<details>

<summary>What is the maximum duration for a voice call?</summary>

The call duration can be set to 5s, 180s, or 600s. After this limit, the call will automatically end, regardless of the conversation state.

</details>

<details>

<summary>Where can I access the Voice Bot call logs or transcripts?</summary>

You can find all transcribed conversations under the Inbox section of your BotPenguin dashboard.

</details>

***

### Support Prompt

If you still have questions for our team, write to us at [**support@botpenguin.com**](mailto:support@botpenguin.com).\
We’ll get back to you within 48 hours.


---

# Agent Instructions: 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:

```
GET https://help.botpenguin.com/ai-agents/creating-ai-agents-in-botpenguin/agent-settings/voice-bot.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.