I am trying to create a workflow that will automate my inbound support calls. I am using Telnyx for calls, Eleven Labs for the voice and LLM for understanding and responding to conversation.
I have already configured the receiving of the call, answering it and playing the basic message using the below mentioned JSON.
{
"voice": "female",
"language": "en-US",
"payload_type": "text",
"payload": "Hello from Telnyx. Please tell me how I can help you.",
"gather_options": {
"end_on_silence_timeout": true,
"silence_threshold": 3000
},
"webhook_url": "YOUR_VOICEBOT_BRAIN_PRODUCTION_WEBHOOK_URL"
}
This worked successfully, but I am failing to play the file generated by Eleven Labs to make it work. I want to specifically use this infrastructre to minimize my costing and not got for conversational AI.
Can anyone help me where I am going wrong in making it work using the code below?
{
"voice": "female",
"language": "en-US",
"payload_type": "ssml",
"payload": "<speak><audio src=\"data:audio/mp3;base64,{{ $('Generate Welcome Speech').item.binary.data.toString('base64') }}\"></audio></speak>",
"gather_options": {
"end_on_silence_timeout": true,
"silence_threshold": 3000
},
"webhook_url": "YOUR_VOICEBOT_BRAIN_PRODUCTION_WEBHOOK_URL"
}
Thanks for your responses in advance.