Riddi

Riddi turns web articles into natural-sounding speech using local AI models that run directly in your browser. No cloud TTS service. No API keys. No article text sent to a server for speech generation. Use Riddi when you want to listen instead of read: long articles, documentation, essays, research, tutorials, or any page where your eyes need a break. What Riddi does: • Reads clean article text aloud Riddi extracts the main content from a page and skips common clutter like navigation, ads, and sidebars. • Runs locally Speech is generated on your device with bundled neural TTS models using ONNX Runtime. WebGPU is used when available, with WebAssembly fallback. • Supports multilingual reading Riddi uses Supertonic 3 with support for 31 languages. It can detect the page language automatically, or you can choose one manually. • Lets you choose what to read Use full-page article playback, or select a specific text block when you only want part of the page. • Keeps your place visually The current paragraph is highlighted while speech plays, with word-level highlighting and smooth scrolling. • Includes built-in voices Choose from 10 bundled voice styles, adjust playback speed, and tune quality steps. • Stays out of the way Use the floating player on the page, or disable it and control playback from the extension popup. Riddi is built for private, local, distraction-free listening.