ReadX Text To Speech

Text to Speech Reader – Read Web Pages Aloud (TTS)

Text to Speech (TTS) reader that reads web pages aloud. Listen to articles, blogs, and selected text using Microsoft voices, offline system voices, or AI-generated voices.

Features
Read web pages aloud instantly from any point on the page
Select text to start reading or use keyboard shortcuts
Replace words and abbreviations to fix pronunciations
Works online and offline
Custom play/pause keys, highlight color, and reading speed
Displays reading progress, estimated time, and percent read
Voice Support
Microsoft natural text-to-speech voices
System voices (offline support)
OpenAI-based Text-to-Speech API
Local AI-generated voices with Piper ONNX

How to Use
Click anywhere on a page to start reading from that point, or select text and press the play shortcut. Options allow disabling click-to-read and customizing controls for your workflow.

Supports OpenAI-based Text-to-Speech API endpoints
Added support for Piper ONNX voices

New in Version 1.4.0:
- Added support for non-DRM EPUB and MOBI files. Access it by clicking the EPUB / MOBI / Text reader button in the Extension popup.

New in Version 1.3.3:
- Added Tesseract OCR (Optical Character Recognition) to read text or books from images. Runs entirely locally on your device.

New in Version 1.3.2:
- Kokoro can read phonemes as well as plain text. (v 1.3.1)
- Fixed major bug I introduced in previous version of Kokoro phonemization.
- Piper voices now read the bitrate and language settings from the JSON file.
- Minor other bug fixes.

New in Version 1.2.9:
- On secure pages, when using non-system voices, setting #playInPopup:no will open the player in a new tab rather than a popup.
- Using {Lcase}, allows you to convert words with 5 or more uppercase letters in this example to lowercase:
RegEx=>\b\p{Lu}{5,}\b=>{Lcase}=>gu

New in Version 1.2.6:
- Added support for Azure voices, remote parameters now accept XML when the first character is a '<' symbol.
Note: Kokoro WebGPU will default to CPU and run relatively slow compared to GPU, if WebGPU flag isn't enabled.

New in Version 1.2.5:
- Restored support for reading contenteditable elements.
- Added #filter:contenteditable=true to disable reading these elements if desired.
- Introduced Kokoro WebGPU: prompts a one-time 300MB TTS model download. When
selected, it enables reading selected text in a new window via the option menu.
- Click the Edit selected voice button to bring up a list of 28 available voices.

New in 1.2.4
- Fixed hotkey case sensitivity
- Added checking to prevent edit boxes from accidentally triggering read events.
- Play/Pause key now works in iframes

New in 1.2.2 - 1.2.3
- Added offline support for Piper ONNX voices. ( press edit selected voice for its options )
- Improved GUI for OpenAI-based Text-to-Speech API endpoint and custom voices
To add more you'd have to download the onnx and json files from the extensions add new voice dialog
- v 1.2.3 minor bug fixes, added more language support.

New In 1.2.1
- updated color picker for highlight
when "highlight text only" is enabled, the background color is used on light fonts
- added #filter to ignore entire element blocks when reading,
useful for skipping sections like advertisements, menus or buttons.
- added option for monitoring the page for changes while reading.
#detectPageChanges:yes

New in 1.2.0
- Added color picker for highlights, right clicking on it changes background color. Clicking on the preview color resets to default.
- Upped maximum speed to 3x for local voices.
- Added filter to lowercase capitalized text:
This changes words with 5 or more capital letters to lowercase.
RegEx=>\b[A-Z]{5,}\b=>{Lcase}=>gm

New in 1.1.8
- Added support for OpenAI-based Text-to-Speech API endpoint.

New Features of Version 1.0.2 (iframes not supported)
- Automation tab settings
- Spacebar to play from selected location
- Configurable to automatically click on [next page] when finished reading current one (requires refresh)
* This works by getting element class name, if multiple elements have the same class name on page, this won't work.

New Features of Version 1.0.9
- added domain specific settings
- optimized code for even faster loading

New in Versions 1.1.0 - 1.1.4, 1.1.9
- Minor bug fixes

New in Version 1.1.5 - 1.1.8
-Added an optional Shift key modifier for onclick. Although a better alternative is to set the #playPauseShortcutKey and use the mouse to select the text. Note: The default Play/Pause shortcut is the Spacebar, which will be replaced by the custom key you set. Ensure the 'Spacebar to play' is enabled for the shortcut keys to work.

- Added Custom voice server support. Details available in extension options.
- Disclaimer: This feature is still a work in progress.

Note:
Some features may work differently inside iframes or Shadow DOM elements.
Some pages may have text marked as invisible for text readers. For ReadX to see that text add the following line to the replace textbox.
#readAriaHidden:yes