Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

iOS Example

The following example uses the iOS SDK for Amazon Polly to read the specified text using a voice selected from a list of voices.

The code shown here covers the major tasks but does not handle errors. For the complete code, see Amazon Mobile SDK for iOS Amazon Polly demo.

Initialize

// Region of Amazon Polly.
let AwsRegion = AWSRegionType.usEast1
 
// Cognito pool ID. Pool needs to be unauthenticated pool with
// Amazon Polly permissions.
let CognitoIdentityPoolId = "YourCognitoIdentityPoolId"
 
// Initialize the Amazon Cognito credentials provider.
let credentialProvider = AWSCognitoCredentialsProvider(regionType: AwsRegion, identityPoolId: CognitoIdentityPoolId)

// Create an audio player
var audioPlayer = AVPlayer()

Get List of Available Voices

// Use the configuration as default
AWSServiceManager.default().defaultServiceConfiguration = configuration

// Get all the voices (no parameters specified in input) from Amazon Polly
// This creates an async task.
let task = AWSPolly.default().describeVoices(AWSPollyDescribeVoicesInput())
 
// When the request is done, asynchronously do the following block
// (we ignore all the errors, but in a real-world scenario they need
// to be handled)
task.continue(successBlock: { (awsTask: AWSTask) -> Any? in
      // awsTask.result is an instance of AWSPollyDescribeVoicesOutput in
      // case of the "describeVoices" method
      let voices = (awsTask.result! as AWSPollyDescribeVoicesOutput).voices
 
      return nil
})

Synthesize Speech

// First, Amazon Polly requires an input, which we need to prepare.
// Again, we ignore the errors, however this should be handled in
// real applications. Here we are using the URL Builder Request,
// since in order to make the synthesis quicker we will pass the
// presigned URL to the system audio player.
let input = AWSPollySynthesizeSpeechURLBuilderRequest()

// Text to synthesize
input.text = "Sample text"

// We expect the output in MP3 format
input.outputFormat = AWSPollyOutputFormat.mp3

// Choose the voice ID
input.voiceId = AWSPollyVoiceId.joanna

// Create an task to synthesize speech using the given synthesis input
let builder = AWSPollySynthesizeSpeechURLBuilder.default().getPreSignedURL(input)

// Request the URL for synthesis result
builder.continueOnSuccessWith(block: { (awsTask: AWSTask<NSURL>) -> Any? in
	// The result of getPresignedURL task is NSURL.
	// Again, we ignore the errors in the example.
	let url = awsTask.result!

	// Try playing the data using the system AVAudioPlayer
	self.audioPlayer.replaceCurrentItem(with: AVPlayerItem(url: url as URL))
	self.audioPlayer.play()

	return nil
})