Voice Anywhere

On-device dictation for any text field. Whisper runs locally — your voice never leaves this computer.

Voice Anywhere lets you hold a key, speak, and type into whatever field is focused — search boxes, emails, tickets, docs, and almost any input on the web. The difference is where the work happens. Chrome’s built-in speech tools and most dictation extensions send audio to a cloud API. Voice Anywhere runs OpenAI’s Whisper model inside the extension. Microphone audio is transcribed on your computer. It is not uploaded. There is no account. Why people install it • Local first. Default English model is bundled. After install, it works offline. • Safety first. No speech vendor, no training on your voice, no analytics SDK. • Works where cloud dictation fails. Sites that block Web Speech still get text at the cursor. • Any field. Search boxes, comboboxes, contenteditable editors, and many shadow-DOM inputs. • Hold or toggle. Control, Alt, Caps Lock, F8, or a key you choose. • Voice commands. “new line”, “period”, “comma”, “scratch that”. • Per-site off switch. Disable it on a page from the popup. • History stays here. Recent transcripts live in your Chrome profile so you can copy or undo. How to use it 1. Click the extension icon and allow the microphone (Chrome cannot ask inside the popup). 2. Click a text field. 3. Hold Control, speak, release. Text is inserted at the cursor. Open Settings to change the hotkey, switch to toggle mode, or pick a larger model. Larger models are optional downloads from Hugging Face. Audio is still transcribed on-device after that file is cached. Google Docs uses a custom editor, so insert there is best-effort. The extension cannot run on chrome:// pages or the Chrome Web Store.