Utter

About

Utter is a lightweight Chrome extension that brings voice-to-text input to any webpage. Simply press a keyboard shortcut or hold your custom key combination, speak naturally, and watch your words appear in real-time in any text field.

**Key Features:**

• **Universal Voice Input** — Works in any text field on any website. Write emails, fill forms, compose documents, or chat — all by speaking.

• **AI-Powered Text Refinement** — After transcribing, press Alt+R to automatically improve your text using Chrome's built-in AI (Gemini Nano). Remove "ums" and "uhs", fix grammar, or change the tone to be more formal or friendly. Choose from preset styles or create custom refinement prompts. Both original and refined versions are saved in your history.

• **Two Activation Modes:**
  - **Toggle Mode** — Press Ctrl+Shift+U (Cmd+Shift+U on Mac) to start listening, press again to stop
  - **Push-to-Talk** — Hold your custom key combination while speaking, release to stop. Perfect for quick voice inputs.

• **Voice Input History** — Access your transcription history in the side panel. Review past inputs with timestamps and source URLs, and easily delete entries you no longer need.

• **Audio Recording & Playback** — Automatically records audio alongside your transcriptions. Listen back to your recordings with an inline audio visualizer, verify transcription accuracy, and download both text and audio files.

• **Microphone Selection** — Choose which microphone to use in the options page. Great for users with multiple audio devices.

• **Visual Feedback** — A small floating indicator shows when Utter is listening and displays your speech in real-time as it's transcribed.

• **Audio Feedback** — Optional sounds let you know when voice recognition starts and stops.

• **Privacy Focused** — All speech processing uses Chrome's built-in Web Speech API. Text refinement uses Chrome's built-in AI (Gemini Nano), which runs entirely on your device. Your voice data is processed by Google's speech recognition service (the same service used by Google Search voice input) but is never stored by Utter. No account required, no data collection.