About this plugin
AI Text-to-Speech using AWS Polly creates audio versions of WordPress posts with AWS Polly voices.
This is an independent plugin by iTRON. It is not affiliated with or endorsed by Amazon, AWS, or Amazon Polly.
Development and source code: GitHub repository
Key features:
- Generate audio for supported post types.
- Queue audio generation in the background with WP-Cron.
- Store audio locally or in Amazon S3.
- Optionally use Amazon CloudFront for delivery.
- Control voice, sample rate, autoplay, player label, and download availability.
Configuration
The complete setup instructions are maintained in the configuration guide on GitHub. Please follow that guide to configure AWS credentials, Amazon Polly, and optional Amazon S3 or CloudFront delivery.
External services
This plugin connects to services provided by Amazon Web Services, Inc. (AWS) to generate, store, and optionally deliver audio files after an administrator configures the corresponding functionality. AWS charges may apply.
Previously generated S3 objects may also require deletion requests after generation is disabled or storage settings change.
Amazon Polly
Amazon Polly is used to convert post content into audio files.
Data sent when audio is generated: the enabled portions of the post title, excerpt, and content prepared as text or SSML for speech synthesis, plus the selected voice, engine, sample rate, output format, configured lexicon names, and AWS region. The configured AWS credentials are used on the server to authenticate and sign requests; the secret access key is not sent as content.
Generation can be queued by saving an audio-enabled post or by the Generate Audio bulk action, and runs through WP-Cron. While generation is enabled and credentials are configured, settings, editor, and generation flows may request the available voice catalog from Polly. Catalog requests include the region and pagination parameters, not post content; results are cached.
com/polly/
com/service-terms/ (AWS Service Terms, including the AWS Machine Learning and Artificial Intelligence Services section that covers Amazon Polly)
com/privacy/
com/compliance/data-privacy-faq/
Amazon S3
Amazon Simple Storage Service (Amazon S3) stores generated audio when S3 storage is enabled. Existing S3 audio can still need cleanup after switching to another storage location or disabling generation.
Data sent on upload: the resulting MP3 audio file, its object key, content type, selected bucket, and AWS region. Access validation may check the configured bucket. Replacing or deleting audio sends its recorded bucket, region, and object key in a deletion request.
The configured AWS credentials are used on the server to authenticate and sign requests; the secret access key is not sent as content. com.
When S3 storage is enabled and CloudFront is not configured, visitors also download the generated audio files directly from your Amazon S3 bucket when they load a page with audio. Those requests include the audio file URL and standard browser request data such as the visitor IP address and user agent.
Before displaying a remote audio player, the WordPress server may send a cached availability check (HTTP HEAD) to the recorded audio URL. This discloses the audio URL and server request information, not visitor IP addresses or post text.
com/s3/
com/service-terms/ (AWS Service Terms for AWS services)
com/privacy/
com/compliance/data-privacy-faq/
Amazon CloudFront
Amazon CloudFront is used only when you configure a CloudFront domain for audio delivery.
Data sent when visitors load a page with audio: requests for the generated audio files are served through your configured CloudFront distribution. Those requests include the audio file URL and standard browser request data such as the visitor IP address and user agent.
WordPress may also check that the audio URL is available using a cached HTTP HEAD request from the server, as described above for S3.
com/cloudfront/
com/service-terms/ (AWS Service Terms, including the Amazon CloudFront section)
com/privacy/
com/compliance/data-privacy-faq/