Voice input (dictation).

LADLE supports dictation into the composer using the Web Speech API built into your browser. Click the microphone icon (visible in the composer when your browser supports it), grant microphone permission the first time, and speak. The transcript appears in the composer as you talk. Click the icon again to stop.

**Browser support:** - Chrome, Edge, Opera (desktop): full support - Chrome (Android): full support - Safari (macOS 14+, iOS 17+): full support - Firefox: NOT supported (Mozilla has not shipped SpeechRecognition). The mic icon is hidden entirely.

**Privacy:** The audio stream never touches LADLE's servers. It flows: microphone → browser → your OS's speech service (Google's on Chrome/Android, Apple's on Safari/iOS, Microsoft's on Edge) → transcript → composer. Only the resulting text is sent to LADLE when you hit send. Whether the transcription itself happens locally or in the cloud depends on your OS's dictation settings — this is not our decision to make; the Web Speech API's architecture routes through the OS layer.

**Language:** Transcription uses your browser's UI language by default. Change it in your OS settings or (on Chrome) via chrome://settings/languages. Multi-language dictation (mid-sentence code-switch) works best on Chrome with "auto detect language" enabled in the speech settings.

**Known limits:** - Silence detection ends the session after ~2-3 seconds of no speech (browser-controlled). - Long-form dictation may cut off — restart the mic to continue; your composed text is preserved. - Punctuation is inferred by the recognizer, not perfect. Say "period" or "comma" to force them. - Special characters and code are usually easier to type.

**If it's not working:** Check that microphone permission is granted for our domain in your browser's site settings. If the mic icon isn't visible, your browser doesn't support the API — try Chrome or Safari.