🚨 Issues Vapi x VF Integration 🚨
The new Vapi integration with Voiceflow looks promising but I ran into some issues:
I have been trying out the new Vapi integration with Voiceflow to check and see if I can use my Voiceflow agent as a backend and the Vapi voice platform as a frontend. Basically, in my Voiceflow assistant I am using an API call to an AI assistant that I am running on a server to process the user’s input and return a valid response.
However, with this new integration, somehow when I ask the assistant to mention all the products that a store has, I can see in the console in my server that the assistant outputs a response of about 10 products (which is correct) but somehow when the voice assistant is turning this text response into voice output, it only outputs the first and latter part of the response. So, for example, it will say something like:
“Sure, we have a whole list of products. …… Product number 10 is a beach bag. Would you like to know more about any of these products?”
So it will skip the 9 other products while still for some reason mentioning the last product.
Has any one of you have had issues with this? Because when I use the exact same test in the Voiceflow chat window, then I do get a full response. Can this be because I am using markdown formatting and letting the assistant output responses in a JSON format?
Hope someone can help me out with this
I have been trying out the new Vapi integration with Voiceflow to check and see if I can use my Voiceflow agent as a backend and the Vapi voice platform as a frontend. Basically, in my Voiceflow assistant I am using an API call to an AI assistant that I am running on a server to process the user’s input and return a valid response.
However, with this new integration, somehow when I ask the assistant to mention all the products that a store has, I can see in the console in my server that the assistant outputs a response of about 10 products (which is correct) but somehow when the voice assistant is turning this text response into voice output, it only outputs the first and latter part of the response. So, for example, it will say something like:
“Sure, we have a whole list of products. …… Product number 10 is a beach bag. Would you like to know more about any of these products?”
So it will skip the 9 other products while still for some reason mentioning the last product.
Has any one of you have had issues with this? Because when I use the exact same test in the Voiceflow chat window, then I do get a full response. Can this be because I am using markdown formatting and letting the assistant output responses in a JSON format?
Hope someone can help me out with this
