Grok Voice Reader

Reads a new Grok reply aloud in the tab that is open.

Short description Reads a new Grok reply out loud in the open tab, using the speech voices already on your computer. Single purpose Reads Grok replies aloud in the tab that is open, using speech voices on the computer, including an optional Piper voice the user downloads. Detailed description Grok Voice Reader reads a new Grok reply out loud in the tab you have open. It works in Chrome. Edge and Brave can install the same listing. A new reply is read after it stops changing. Your own messages stay quiet. Each bot can have its own voice and its own on or off switch. Those choices stay in the browser. The reply being read has an outline and a Play/Pause button in the corner. The outline is green while reading is on, and light orange while paused. Click another reply to read that one from the start. Pause keeps your place in the same reply. While paused, a new reply is outlined and stays quiet until you press Play. The sentence being spoken is tinted so you can follow it. Voices are the ones already installed on the computer. If those are few, you can download a free Piper voice. The speech runtime is part of the extension. The voice model is a data file stored in this browser. The extension does not send reply text anywhere. Permission justifications storage: Saves your voice, rate, per-bot on/off switch, and play or pause mode locally. Nothing is uploaded. offscreen: Runs Piper speech in a hidden extension document so the bundled WebAssembly can turn a downloaded voice model into sound. It loads no remote code. grok.com, x.ai, grok.x.ai, cursor.com (and their subdomains): The content script reads the Grok chat on these sites so it can speak replies. api.github.com: Unpacked and GitHub installs check once a day for a newer version. Store installs skip this. No reply text is sent. huggingface.co and *.hf.co: Downloads the Piper voice model the user picks from rhasspy/piper-voices. It's model data, not code.