Mews Reader: Private Full-Page TTS

Listen to text, articles, blogs, and web pages without giving up your privacy. Kitten TTS WebGPU runs a lightweight neural text-to-speech engine entirely on your device, powered by your graphics card through WebGPU. WHY THIS ONE IS DIFFERENT - It's free and nothing you read gets sent to a server or API. - No cloud API. No telemetry. No logins. No ads. Just mews. Requirements: 1. [Required] A WEBGPU-compatible dGPU or iGPU - check chrome://gpu 2. [Required] BROWSER settings set to enable "graphics acceleration" - check chrome://settings/system 3. [Recommended] For Hybrid setups: set SYSTEM settings to prefer High Performance GPU for your browser - e.g.: Settings: Display > graphics > "use High Performance GPU" for Chrome - A dedicated GPU is *strongly* recommended. Integrated graphics (iGPU) works, but a short article can take 30+ seconds to synthesize. - On hybrid laptops (dGPU +iGPU), chrome usually defaults to the iGPU so you must set this manually Key Features: • Hardware-Accelerated Synthesis: Generates speech directly on your local GPU via WebGPU, ensuring responsive audio with zero remote API requests. • Smart Page Reading: Automatically extracts clean article prose from blogs, news, and technical documentation while skipping navigation bars, ads, and footers. • Natural Neural Voices: Choose from 8 expressive voices across male and female profiles (Jasper, Bruno, Hugo, Leo, Kiki, Luna, Rosie, Bella). • Configurable Quality Tiers: Select between lightweight, balanced, or high-definition models to match your device's hardware capabilities. • Instant Offline Caching: The audio you generate is cached locally for instant replay, and you can export full recordings directly as 24kHz WAV audio files. • Seamless Integration: Trigger reading with Alt+Shift+A, right-click any highlighted paragraph, or open the side panel with Alt+Shift+K for full control. Ways to Use: 1. Highlight text on any page, right-click, and choose "Read Selected Text". 2. Press Alt+Shift+A on any article to automatically scan and listen to the page. 3. Press Alt+Shift+K to open the side panel anytime to adjust speed, switch voices, or generate audio from typed custom text. 4. Click "save audio" button after generation is finished to export the most recent wav 5. If anything stops working right or you want to stop speech synthesis, hit the "engine reset" 6. If you are on a slow cpu or iGPU, try toggling the "render all audio before playback" button in the sidepanel. Attribution: Powered by the KittenTTS v0.8 by KittenML (Apache-2.0) Running fully on-device via WebGPU as converted WSGL shaders hand-written by Svenfold With primary article extraction by Mozilla's Readability. Original Espeak-ng dictionary & phonemizer derived from Xenova. Privacy by Design: • No user data, text, or browsing history is ever collected, stored remotely, or transmitted. • Works completely offline with self-contained model weights. • Open-source under GPL-3.0: https://github.com/cinnavun/kitten-tts-webgpu-chrome Support & Feedback: If you have questions, feedback, or need assistance, visit our GitHub Issues page or contact me at cinnavun@icloud.com. This is a goodwill project, not a business. It is free, the code is public, and there is nothing to upsell. If something isn't working right on your system, I would appreciate your feedback and patience as I improve on it. https://github.com/Cinnavun/kitten-tts-webgpu-chrome