TTSForX

ttsforx-preview.mp4

TTSForX reads posts on X aloud in your browser. It uses the official Supertonic 3 model through WebGPU. Speech, generated audio, and caption timing stay on your device.

What it does

• Includes all ten Supertonic 3 voices.
• Reads regular posts, long posts, Notes, and Articles.
• Lets you change the voice, speed, volume, and generation quality.
• Supports pause, resume, restart, voice previews, and switching voices while a post is playing.
• Can start at any word in a post.
• Shows karaoke captions that follow the generated audio.

The caption timing comes from audio alignment, not a character-count estimate. Whisper Tiny aligns the known words to the generated PCM, and the captions follow the browser's audio clock during playback.

Install

1. Download tts-for-x-0.2.0-chrome.zip from the latest release.
2. Extract the ZIP.
3. Open chrome://extensions.
4. Turn on Developer mode.
5. Click Load unpacked and choose the extracted folder.

The first use downloads the pinned model files from Hugging Face. Chrome caches them for later sessions. You need a Chromium browser with WebGPU support.

Build from source

npm install
npm run typecheck
npm test
npm run build

Load .output/chrome-mv3 from chrome://extensions. For local development, run npm run dev.

The build check verifies the extension policy, packaged WebAssembly runtime, icons, and WebGPU-only inference setup.

Privacy and licenses

The privacy policy explains what the extension reads and stores. You can report problems through GitHub Issues or support the project on Ko-fi.

The extension source uses the MIT License. The Supertonic 3 model uses OpenRAIL-M. Other dependency licenses are listed in THIRD_PARTY_NOTICES.txt.