Introduction

The **"TTS" (Text-to-Speech)** component allows you to transform text into audio within **Voice-type routing points**. Its main function is to dynamically generate and play audio from text or flow variables. Additionally, it includes a recording mode that allows you to save the generated audio in a variable for reuse in different parts of the diagram, optimizing time and costs.

This component is located in the **"Cognitive"** group, and its configuration differs from the TTS used in **Chat routing**. Therefore, this article explains exclusively its functionality in **Voice Routing**.

Configuration

The **TTS component** has two operating modes: **PLAY** and **REC**, each designed to meet different needs within the flow.

Configure TTS in PLAY Mode

**PLAY mode** immediately plays the audio generated from the text defined in the component. Steps to configure it:

• In the side configuration window, under **"TTS Type"**, select **"PLAY"**.
• In the **"Voice"** field, choose the voice model you want to use.
  • The language is indicated at the beginning of the voice name.
  • Voices labeled as **"neural"** are the most advanced and realistic.
• In the **"Text to Convert or Variable"** field, enter:
  • The text the customer will hear, and/or a variable whose value will be converted to audio.
• Use the ▶ (Play) icon to preview the generated audio.
• (Optional) Enable the **"Interrupt TTS when customer voice is detected"** checkbox if you want to allow the customer to interrupt the audio by speaking.
• Click **"Save TTS Configuration"**.

Configure TTS in REC Mode

**REC mode** converts text to speech and also saves the resulting audio in a variable, allowing it to be played multiple times from other components (e.g., with a **Play Audio**), optimizing TTS generation and reducing billing costs.

Steps to configure it:

• Under **"TTS Type"**, select **"REC"**.
• In the **"Variable to Save TTS"** field, enter the name of the variable where the audio will be stored.
  • Example: $tts_welcome
• In the **"Voice"** field, choose the type of voice to use.
  • Voices labeled as **"neural"** are the most advanced and realistic.
• In the **"Text to Convert or Variable"** field, enter the text and/or variable you want to convert to audio.
• Play the result using the ▶ icon.
• Click **"Save TTS Configuration"**.

Create a Custom Voice with wolkvox TTS

The component includes an additional feature: creating **custom voices** through an audio training process.

1. Click the button **"Would you like to have a custom voice?"** located next to the voice selection field.
2. Fill in the following fields:
   • Language.
   • Full name of the voice actor.
   • Name of the custom voice.
   • Text to read (must exactly match what will be recorded).
   • Record the audio using the buttons:
     • ● Record.
     • ■ Stop.
     • ▶ Play.
   • When you are satisfied, click **"Train wolkvox TTS"**.
3. Note: The initial training has a one-time cost, followed by a monthly fee for storage and usage.