Speech-to-text and text-to-speech — voice interface for chat, forms, and AI responses
The Workbench includes speech-to-text (STT) for voice input. Use the microphone button on chat screens, form fields, and the AI query bar to speak instead of type. STT uses your configured AI provider or the browser's built-in Web Speech API as a fallback.
AI responses can be read aloud via text-to-speech (TTS). Each AI chat message and recipe suggestion has a play button. TTS uses your configured AI provider's TTS endpoint, or the browser's built-in Speech Synthesis API.
Both STT and TTS support multiple languages. The audio language follows your UI language preference. You can override the audio language independently in Profile → Settings — for example, English UI with Thai speech input. The globe icon on the login screen sets the initial audio language.
Audio settings are stored per user: STT language, TTS language, TTS voice (where the provider offers multiple voices), speech speed, and whether to auto-play TTS responses. The admin can disable audio features entirely via Admin → Setup → Options → Audio I/O.