Voice to Text for ChatGPT

Speak instead of typing. Auto-send on silence, push-to-talk, floating mic panel, and more.

Voice to Text for ChatGPT **Short description** Speak instead of typing. Hands-free dictation with auto-send on silence, push-to-talk, and a floating microphone panel. **Full description** Turn your voice into text on ChatGPT. This extension adds a floating microphone panel to the ChatGPT interface, letting you dictate messages hands-free. No extra apps or cloud services required — everything runs locally in your browser. Key features: - Real-time speech-to-text in the ChatGPT input field - Auto-send after a configurable silence delay - Push-to-Talk (walkie-talkie) mode — hold a key, speak, release to stop - Collapsible floating panel with drag-and-drop positioning - Adjustable ChatGPT content width - Light / Dark / System theme selector - Voice punctuation commands in your dictation language (~30 languages) - Word replacements in any language - Support for multiple recognition languages Your voice data never leaves your device. All processing happens inside the browser using the built-in Web Speech API. **What’s New** **Version 2.5** - Major performance optimization: smoother dragging, faster speech recognition rendering, and reduced main-thread work during ChatGPT streaming. - Theme selector: Light / Dark / System — dark mode is fully restyled and follows your OS live. - Voice punctuation commands now work in your dictation language (~30 languages) — say "comma" or "запятая" and get "," with no stray space. - Word replacements now work in any language and ignore case. - Language picker with instant search and a scrollable list. - Content width slider extended to 300% and made resilient to ChatGPT layout changes. - Extension popup: Center Mic, Center Panel, and Open Settings buttons. - Floating controls reliably keep their position — offscreen positions are recovered and clamped to the viewport. - Removed the obsolete "Auto continue" option. **Version 2.4** - Settings modal visually refreshed with a modern header, card-based layout, two-column language grid, and a compact footer. - Smoother interim speech recognition spacing — spaces appear as words are recognized instead of only after finalization. - Security and reliability improvements: reduced extension resource exposure and fixed message handler leaks.