Voice interface

General
1

How can I bring synthflow with voice interface into a PA?

On that topic, how can I control usage of the voice API to limit my cost?

1 Like
2

Yes, been waiting for this option as well.

3

We are currently working on a voice-integrated deployment for your Pickaxes. How would you guys best like to utilize it?

3 Likes
4

Thanks for the response back! at this moment I just need the ability for the end-user to be able to hear back the response as a voice as well ( maybe automatically ,or they press a button).

5

One use case would be for mock interviews were the candidate uploads the initial information such as resume and job ad and then based on the prompt the pickaxe would do a mock interview using voice interaction. the same for sallary negotiations etc. On top of that the benefit using a pickaxe would be to be able to control the number of tokens used to a degree because that could become quite token and cost intensive which with outside tools would present a challenge. as far as the integration itself goes, it would be great to have access to the elevenlabs voices as well as to the real-time voice interface from OpenAI in a way synthflow is currently integrated it

1 Like
6

I have the exact same use case. +1

7

Working for a migration agency and they have a similar use case. They want the clients to have a preliminary visa discovery session with a voicebot.

8

The voice feature is highly relevant for certain use cases. We have clients who want to interact with their documents and receive audio responses to their queries. Of course, this should be optional, allowing users to choose whether they want their voice input to be converted to text and, once the output is generated, to click a button to have the response read aloud. Additionally, offering a seamless conversational mode, similar to ChatGPT’s current functionality, could be another valuable option.

Any of these features will require a clear understanding of token costs and other associated expenses, as we know voice-based interactions are more resource-intensive than simple text outputs. Therefore, it might be advisable to include an option when creating each Pickaxe to enable or disable the respective voice services based on the user’s preferences and needs.

3 Likes