So i wanted to add a custom voice with smallest ai but i guess the api structure is different. So my question is: Is there an woraround that could work? smallest.ai:

curl --request POST \
  --url https://waves-api.smallest.ai/api/v1/{model}/get_speech \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "voice_id": "emily",
  "text": "<string>",
  "speed": 1,
  "sample_rate": 24000,
  "add_wav_header": true
}'

https://cdn.discordapp.com/attachments/1321083349967704114/1321083350248587346/image.png?ex=676bf23f&is=676aa0bf&hm=338b5ef8a52c5fb757d1473189afcc62892317923b54ce330cdd9b727003bc4a& https://cdn.discordapp.com/attachments/1321083349967704114/1321083350701838336/image.png?ex=676bf23f&is=676aa0bf&hm=1ffb8bcef851092f3a0fa2df579f7dd5e4263cf14889abbf7e331b7ed5f8c8ba&

@Manan Khurana Certainly! To integrate a custom voice from a different API like smallest.ai with VAPI, you would typically need to create a wrapper or an adapter that translates the API requests and responses between the two services. Here's a general approach you could take: 1. **Identify the differences**: Understand the request and response formats of both APIs to identify the differences. 2. **Create a translation layer**: Develop a middleware that can take the VAPI request format and convert it into the smallest.ai request format. 3. **Handle authentication**: Ensure that your middleware can securely store and use the authorization tokens required by smallest.ai. 4. **Convert responses**: After receiving the response from smallest.ai, convert it back into the format expected by VAPI if necessary. 5. **Error handling**: Implement robust error handling to manage any issues that arise during the API calls. If you need assistance with the specifics of setting up this integration or if you have any other questions, please let me know!

@Manan Khurana can you please rephrase your issue description I am unable to understand it properly what you wanted to achieve with smallest_ai and Vapi.

Hey we figured a workaround with request transformer with make.com. So it's resolved. Thank you @User