deep-jade•16mo ago
Hi People,
Still fairly new to VF and learning the ropes a little. Can anyone tell me how to create speech to voice, I want the AI to respond but using a voice rather than just text. I have it set-up to read documents/data I've uploaded but at the minute it's jus text responses, looking to convert this in to conversational.
Thanks in advance!
10 Replies
To create a speech response in Voiceflow, you can use the Speak Step to enable your AI assistant to respond using text-to-speech (TTS). Here’s a step-by-step guide to help you set it up:
1. Add a Speak Step:
- In your Voiceflow project, drag and drop the Speak Step from the steps menu into your flow.
2. Configure the Speak Step:
- Click on the Speak Step to open its configuration panel.
- In the text input field, type the response you want your assistant to say.
- You can also add audio and SSML (Speech Synthesis Markup Language) effects to enhance the speech.
3. Select a TTS Voice:
- You can choose from supported TTS voices (e.g., Alexa and Microsoft TTS voices).
- Set a default voice for your assistant by clicking the star icon next to the voice or selecting it on the settings page.
4. Preview the Speech:
- Click the Play button in the bottom-left corner of the text input field to preview how the speech will sound.
- You can stop the playback at any time by clicking the Stop button.
5. Using Multiple Voices:
- You can use multiple voices within your assistant alongside the default voice by configuring different Speak Steps with different voices.
By following these steps, you can convert text responses into conversational speech, making your AI assistant more interactive and engaging.
For more detailed information, you can refer to the Voiceflow documentation.
deep-jadeOP•16mo ago
Thanks Tico.... This sounds perfect, except for one part in the steps above I have a question for. In step 2 (second bullet), I want the AI to respond with an answer to a question about uplaoded documents. Does that make sense ?
genetic-orange•16mo ago
You say
speech to voice but i assume you mean text to speech ?deep-jadeOP•16mo ago
@MikeG YES sorry about that Mike, thats exactly what I mean
genetic-orange•16mo ago
I assume that works natively on the speach voicelfow project modality?
I havnt used the speech functionality so not sure, only text chat
if your using the text chat project, you would need to adapt the react-chat package to allow for audio responses
deep-jadeOP•16mo ago
" if your using the text chat project, you would need to adapt the react-chat package to allow for audio responses " so Is TICO wrong here ?
I'm trying to get the AI to respond using voice so yes "text to speech", I want the AI to respond using the data/documents uploaded regarding projects..... Mike, could you show me tutorial on how to do this ? There is a tutorial how to get it to speak BUT it's not what I want, it's half the battle I guess
well, sorry what I meant was is there a tutorial anywhere that shows how to do this.
genetic-orange•16mo ago
If your using the speech modality for your project then i think Tico could be right. If your using the "Chat" modality, then i dont think there is native functionality for it so it would need to be custom built
deep-jadeOP•16mo ago
I've just started again using the voice modality, can we still train the assistant on data/documents uploaded ?
genetic-orange•16mo ago
That i dont know, i have 0 experience with voice modality unfortunately
deep-jadeOP•16mo ago
I figured it out Mike.... you can train them!