A WordPress plugin that converts your posts into audio narrations using AI text-to-speech services, making your content more accessible and engaging for your audience.

This plugin is actively used and maintained by Rest of World. We wrote a blog post with more background information here.

Out of the box: This plugin creates audio files for your WordPress posts using AI text-to-speech and adds a responsive audio player to each post. Visitors can listen to your content instead of (or alongside) reading it.

How it works: The plugin processes your post content (headings and paragraphs) and generates multiple MP3 files that play seamlessly together. Content is split into smaller audio segments to ensure fast loading - your readers don't have to wait for one large file to download before they can start listening.

Content processing: Only heading and paragraph blocks are converted to audio. Other elements (images, galleries, etc.) are skipped, with natural pauses added between sections for a smooth listening experience.

• Requires OpenAI API key and account - You'll need to sign up with OpenAI and pay for TTS usage
• Currently supports OpenAI only - Other TTS services may be added in future versions
• Manual setup required - You'll need to configure API keys and may need to specify information to identify post content
• Storage space - Audio files are stored locally on your server
• Theme compatibility - May require CSS selector adjustment if your theme uses non-standard post markup

• TTS - OpenAI Text-to-Speech integration with multiple voice options
• Fully functional player - Responsive audio player that works on all devices
• Flexible generation - Create narrations manually or automatically on post publish
• Bulk Action - Bulk narration generation for existing content
• Customizable - Custom intro/outro messages for each narration
• Extensibility - Developer-friendly with hooks and filters for customization
• Local hosting - Audio files stored on your server (no third-party dependencies for playback)

• Download the ai-narration-main.zip file and unzip it in your WordPress plugins directory
• Rename it as ai-narration
• In WordPress admin go to Plugins > Add New > Upload Plugin
• Choose the plugin zip and click Install Now
• Click Activate
• Narrations should appear in the menu

Go to the plugin settings page.

• Choose service: Select "OpenAI" (currently the only option)
• Add API key: Paste your OpenAI API key and save
• Select voice: Choose from available voices
• Post content selector: Default is main .entry-content - change this if your theme uses different markup

Click Save Changes.

We recommend that you only enable Auto-Generate on Publish after manually generating a test narration.

• Intro & outro text: Custom text to be read at the beginning or end of each narration. Includes variables for the headline, author and publish date.
• Intro & outro MP3: Paths to custom audio files to be played at the beginning or end of each narration.
• Player placement: Specify where the player should be inserted, as counted by word-count, number of paragraphs, etc.
• Learn More link: A link for an optional 'Learn More' link to appear in the player
• CDN: Serve the audio files from your specified CDN

You can customize certain criteria for posts to be eligible for AI narration.

• Post type
• Tags: Posts with any of the listed tags will be excluded
• Date: Posts before the given date will be excluded
• Word Count: Posts with fewer than the minimum or more than the maximum specified word counts will be excluded

• WordPress 6.0+ (Gutenberg support recommended)
  • Likley works with WP 5+ but untested
• PHP 7.4+
• OpenAI API account with TTS access
• Adequate server storage for audio files

To install and work on AI Narration locally:

cd /path/to/wordpress/wp-content/plugins
git clone https://github.com/row-engineering/ai-narration.git
cd ai-narration

Our roadmap for this plugin is relatively modest. The core functionality is in place. Our future efforts will be centered around stability, incremental changes, and UX improvements.

We are a small team and reviews may not be timely. We also have a narrow focus for this plugin as it actively runs on our site - mostly on security, stability, and improving existing features. We are not looking to add significant features without consultation.

• Service Support
  • Adding support for additional AI TTS services by extending the current system
• WP Admin
  • A cleaner way to abort, and redo, narration generations
• Internationalization
  • Any refactoring to better prepare it for language support
  • Add support for a second language
• Actions and Filters
  • Add new actions or filters at logical/useful points. This may be subjective so use your best judgment.
• Documentation
  • Spelling errors
  • Clearer ways to explain things
  • Addressing any omissions
• Security
  • Please get in touch directly if you find a severe issue and we will prioritize those.

• Keep PRs focused and atomic
• Include comprehensive but relevant descriptions
• Add tests or detailed documentation
• Follow existing code and formatting standards (tabs, not spaces)

Pull requests are encouraged and often welcome. Pick an issue or a feature and help us out!

• Processes specific Gutenberg blocks: heading, paragraph, and seperator blocks
• Skips multimedia, shortcodes, and all other blocks blocks
• Compatible with Classic Editor paragraph content
• Adds natural pauses between sections using separator elements

When narrations are generated, the plugin creates:

• MP3 files: Individual audio segments (audio_01.mp3, audio_02.mp3, etc.)
• JSON metadata file: index.json contains the same data structure as the inline JavaScript (see below)
  • The JSON file can be used for external integrations or caching strategies.
• Directory structure: /wp-content/narrations/YYYY/post-slug/

The plugin inserts inline JavaScript data via a variable in global namespace to each Post that has a narration. This mirrors the data from the JSON file created for each narration. The player relies on this for rendering and playback:

<script id='ai-narration-data'>
window.AINarrationData = {
    "id": 88241,
    "title": "My mom and Dr. DeepSeek",
    "date": "2025-09-02 06:00:00",
    "url": "https:\/\/restofworld.org\/2025\/ai-chatbot-china-sick\/",
    "slug": "ai-chatbot-china-sick",
    "authors": [ "Viola Zhou" ],
    "audio": {
		    "service": "OpenAI TTS",
		    "model": "tts-1",
		    "voice": "shimmer",
		    "created": "1757088590.149598",
		    "total": 8,
		    "duration": [ 203, 247, 257, 240, 246, 245, 235, 41 ],
		    "tracks": [
            "\/wp-content\/narrations\/2025\/ai-chatbot-china-sick\/audio_01.mp3",
            // ... additional tracks
            "\/wp-content\/narrations\/2025\/ai-chatbot-china-sick\/audio_08.mp3"
        ]
    },
    "config": {
        "cdn": "",
        "link": "\/about\/ai-narrations\/",
        "selector": "main .entry-content",
        "position": [false, 0]
    }
}
</script>

Example story: My mom and Dr. DeepSeek

ain_script_src

Override the player JavaScript file:

add_filter('ain_script_src', function ($src, $post) {
  return get_stylesheet_directory_uri() . '/js/narration-script.js';
}, 10, 2);

ain_script_version

Append a query parameter to your custom JS file to bust browser cache after an update:

add_filter('ain_script_version', function ($version, $post) {
	return $version . random_cache_param();
}, 10, 2);

ain_styles_src

Override the player CSS file:

add_filter('ain_styles_src', function ($src, $post) {
  return get_stylesheet_directory_uri() . '/css/narration-styles.css';
}, 10, 2);

ain_styles_version

Append a query parameter to your custom CSS file to bust browser cache after an update:

add_filter('ain_styles_version', function ($version, $post) {
	return $version . random_cache_param();
}, 10, 2);

narration_request

Add your own filters to disqualify a post from AI narration.

add_filter('narration_request', function ( $data = array() ) {
	$post_id = $data['id'];
	$article_flags = get_article_flags($post_id);

	// If post has force-include-ai-narration flag, generate despite other exclusion params
	if ( in_array('force-include-ai-narration', $article_flags) ) {
		error_log("Post $post_id has the force-include-ai-narration flag.");
		return $data;
	}

	// Must not have exclude-from-ai-narration flag
	if ( in_array('exclude-from-ai-narration', $article_flags) ) {
		error_log("Post $post_id has the exclude-from-ai-narration flag.");
		return false;
	}

	// Must use Regular, Medium or Full template
	$template = get_field('row_post_template', $post_id);
	$accepted_templates = array('regular', 'medium', 'full');
	if ( !in_array( $template, $accepted_templates ) ) {
		error_log("Post $post_id is not using one of the supported templates: Regular, Medium or Full.");
		return false;
	}

	return $data;
}, 10, 1);

AI Narration is an open source project with its ongoing development made possible entirely by the engineering team at Rest of World - and potentially others in the future. If you'd like to support these efforts, please consider donation which will also help support our global journalism efforts:

• Donate to Rest of World.

Rest of World is a 501(c)(3) nonprofit organization and donations in the USA are tax-deductible.

View the change log. Return a full URL to your custom stylesheet. Return an empty string to skip loading.