I'm not sure I understand your use case completely. Are you trying to do TTS in your IVR, like to read back a balance or a street address? Or do you want to be able to record new prompts using TTS and then upload them to UCCX?
For the first, UCCX offers some basic TTS-like functionality for numbers and dates, for more advanced stuff like reading back a status or some phrase you can use Nuance. For the second, the Tropo approach is what I would do, but not use Tropo. Amazon, Google, and Microsoft all offer pretty cheap TTS services. Here's some code which takes a CSV and produces MP3s using Amazon Polly. I haven't touched this code in over a year and I slapped it together when I had to do a few hundred prompts quickly, so please don't judge it too hard.
// Load the SDK
const AWS = require('aws-sdk')
var fs = require('fs')
var csv = require('fast-csv');
// Create an Polly client
const Polly = new AWS.Polly({
signatureVersion: 'v4',
region: 'us-east-1'
})
fs.createReadStream('OrderRefills.csv')
.pipe(csv())
.on('data',function(data){
// console.log(data[1]);
let params = {
'Text': data[1],
'OutputFormat': 'mp3',
'VoiceId': 'Kimberly'
}
Polly.synthesizeSpeech(params, (err, stuff)=>{
if (err){
console.log(err.code)
} else if (stuff){
if(stuff.AudioStream instanceof Buffer) {
fs.writeFile("./"+data[0]+".mp3", stuff.AudioStream, function(err){
if(err){
return console.log(err)
}
console.log(data[0]+".mp3 "+"File was saved");
})
}
}
})
})
.on('end',function(data){
console.log('Done');
});
