[Extension] How to transcribe audio messages with the Text to Speech (TTS) extension

1. Introduction

2. Features

3. Installation and configuration

4. Integration with Builder

5. Usage examples

6. Support

Brief description: Convert your chatbot's text messages into audio with OpenAI voices and offer more accessible service on Blip channels.

Introduction

The Text to Speech (TTS) extension converts your chatbot’s text messages into audio using OpenAI voices. The bot starts responding by voice on Blip channels, making the service more accessible and natural for those who prefer to hear the response or have difficulty reading.

For example, when a customer asks about business hours, the bot generates audio with the chosen voice instead of sending a text response. You can also connect the extension to responses from other extensions that generate text, such as ChatGPT or FAQ GPT, and transform dynamic responses into audio.

Features

• Converts any text in the flow into audio using the OpenAI voice model (gpt-4o-mini-tts).

• 10 voices to choose from: Alloy, Ash, Ballad, Coral, Echo, Fable, Nova, Onyx, Sage, and Shimmer.

• Test each voice directly in the Portal with an editable sample text before saving.

• Instructions field to adjust the style and tone of the generated voice.

• Integration block in Builder that accepts fixed text or flow variables.

• Combination with other extensions that generate text responses, such as ChatGPT or FAQ GPT.

Supported on Blip platform channels compatible with audio messages.

Installation and configuration

Attention: You need an OpenAI API key with available credits. The extension validates the key during installation and rejects invalid, expired, or credit-less keys.

1. Activate Text to Speech (TTS) on the Blip Store.

2. Install the extension in the service bot (the chatbot built in Builder) where the conversation flow runs.

3. Open the bot, click the three dots in the upper right corner, and select the Text to Speech (TTS) extension. The configuration screen opens.

4. In the API Key field, enter your OpenAI key (available at https://platform.openai.com/api-keys) and click Save. The extension validates the key and saves the configuration.

5. Choose the voice. In the list, click the play icon to listen to each voice with the sample text, use Change text to edit the test text, and click the selection icon to set the voice. If desired, fill in the Additional instructions field to adjust the style and tone of the voice and use the link See example of a good prompt as a guide. Click Save.

Integration with Builder

To use the extension inside a flow, integrate it with Builder.

1. Install the Chrome extension. On the Integrating with Builder tab of the extension screen, click Install extension.

2. You will be redirected to the Chrome Web Store. Click Add to Chrome and complete the installation. Then refresh Builder.

3. In Builder, click the add block icon (+) and select New integration.

4. Open the Integration tab and select OpenAI Text to Speech.

5. In the Text field, enter the content to be converted into audio. It can be fixed text or a flow variable. Save the block. The defined text is sent to the customer in audio format.

Usage examples

Generate audio from ChatGPT extension responses

Combine the extension with ChatGPT to speak dynamically generated responses.

Note: Install and configure the Text to Speech (TTS) and ChatGPT extensions on the same bot before starting. To configure ChatGPT, see the tutorial or the video.

1. Configure ChatGPT with the bot name, voice tone, creativity, and interaction guidelines with customers.

2. If you want the bot to also understand audio sent by the customer, enable Enable multimodal understanding. This mode allows the chatbot to interpret formats like images, audio, and documents.

3. In the flow, after the ChatGPT block, add a New integration block with OpenAI Text to Speech, as in the previous section. In the Text field, enter the variable that holds the ChatGPT response, for example, {{resultado}}.

Respond in audio only when the customer sends audio

Use output conditions to send audio only in specific scenarios, for example, when the customer sends an audio message.

1. In the extension panel in Builder, click Output conditions.

2. Configure an If block and set the condition with the variable input.content@type.

3. Set the condition for audio content type, using the value that appears in audio messages received on your channel.

4. Click Publish to publish the flow.

5. Test the conversation to check the result.

Support

If you have questions or issues with the extension, contact Wiv support at support@wiv.com.br.

Need more help? Explore our content at Blip Academy or Blip Community, watch tutorials on our YouTube channel, or ask your questions in our support channel 😃