Introduction

The "TTS Audio Response" component allows you to convert written text into an audio message within a chat flow in wolkvox Studio. It is an ideal tool for enhancing multimodal conversational experiences by sending automatically generated audio from text using TTS (Text-to-Speech) technology.

This component is available exclusively for Chat-type routing points. It allows you to define the voice, the text to be converted, and whether the chatbot should continue without waiting for a user response. Additionally, it includes the option to train a custom voice, opening up advanced customization possibilities for brand experiences.

How to Use the "TTS Audio Response" Component in a Chat-Type Routing Point?

Follow these steps to configure it correctly:

1. In the Chat Routing editor, within the Basics group, locate the TTS Audio Response icon and drag it into the flow.
2. In the right configuration panel, you will find the "Voice" field, a dropdown menu where you can choose from the different available voices.
   1. Voices with the word "neural" are the most advanced, with more natural intonation and higher quality.
3. In the "Text to Convert to Audio" field, enter the message that the system will transform into sound.
   1. You can use plain text, dynamic text, or include variables according to your conversational flow.
4. You have the following fields:
   1. The "Continue Chatbot" checkbox indicates whether the flow should continue without waiting for a user response:
      • Checked: The chatbot continues immediately to the next component.
      • Unchecked: The system waits for the user to respond in the chat before proceeding.
   2. Click the "Play" button to listen to a preview of the generated audio and validate that the voice and intonation meet expectations.
   3. Click the "Save" button to apply the configuration and activate the component within the flow.

Train a Custom Voice

1. If you want the audio to use an exclusive voice for your brand or company, click the button "Would you like to have a custom voice?"
2. This will open the training module, where you can record and generate a unique voice for your operation.
   1. How it works: You must record approximately one minute of audio so the system can train the voice.
   2. Cost: It is clarified that the service has an initial training cost and a monthly usage cost.
   3. Intended use: Generate your own voices to convert text to audio in your flows.
   4. "Language": This dropdown menu allows you to select the base language in which the voice will be trained.
   5. "Full Name of the Voice Actor": Here, you enter the real name of the person whose voice is being recorded.
   6. "Voice Name": Allows you to assign an internal name to the custom voice you are creating. Examples:
      • "Support_Voice"
      • "Maria_Voice"
      • "Corporate_Voice_ES"
   7. "Text to Read": In this field, you write the text that the voice actor must read during the recording.
      • Recommendations:
        • Use continuous text that lasts approximately 1 minute.
        • Include a variety of words, numbers, and phrases to improve training quality.
        • This text will be the script that the person reads while recording their voice.
   8. Recording controls: At the bottom, the recording buttons appear.
      • Record.
      • Stop.
      • Play.
      • Upload (Train wolkvox TTS): Sends the recording to the system and starts the automatic voice training.
        • It is essential that the recording is clear, continuous, and without noise to obtain a high-quality TTS voice.
        • The voice will be available as an option within the voice selector of the TTS Audio Response component.
        • The user will be able to convert any text to audio using their new custom voice.