Amazon Polly

Amazon Polly converts text into lifelike speech in a variety of voices and languages.

See Amazon Polly Developer Guide for general guidance using Polly.

API Access

The IPollyClient interface and its implementation TPollyClient provides access to all Polly operations.

Synthesizing Speech

Use SynthesizeSpeech to convert text to audio. The response contains an AudioStream that can be saved to a file:

var
  Client: IPollyClient;
  Request: IPollySynthesizeSpeechRequest;
  Response: IPollySynthesizeSpeechResponse;
  FileStream: TFileStream;
begin
  Client := TPollyClient.Create;

  Request := TPollySynthesizeSpeechRequest.Create;
  Request.Text := 'Hello from Amazon Polly.';
  Request.VoiceId := 'Amy';
  Request.Engine := 'neural';
  Request.OutputFormat := 'mp3';

  Response := Client.SynthesizeSpeech(Request);

  FileStream := TFileStream.Create('output.mp3', fmCreate);
  try
    FileStream.CopyFrom(Response.AudioStream);
  finally
    FileStream.Free;
  end;
end.

Other Operations

• DescribeVoices lists available voices, optionally filtered by language.
• StartSpeechSynthesisTask starts an asynchronous synthesis task that writes output directly to an S3 bucket, useful for longer texts.
• Pronunciation lexicons can be managed with PutLexicon, GetLexicon, ListLexicons, and DeleteLexicon.