FOR THE PATIENTEspañol
Spanish
Translation will appear here.
ONE DEVICE. A SHARED CONVERSATION.
Which language
will your patient speak?
Choose a language to begin English ↔ patient interpretation.
Spoken translation, with captions you can both read.
13 spoken output languagesEnglish-speaking clinician
Chinese uses the API’s zh output. There is no documented dialect selector or separate Cantonese output guarantee. English is included for completeness; English ↔ English may stay silent because the model skips speech already in its output language.
A temporary conversation. Audio goes to OpenAI for translation. This app saves no audio or captions, uses no database or analytics, and clears captions when you end. Use made-up examples only.
YOUR SHARED CONVERSATION
English ↔ Spanish
Connecting…
Allow microphone access when prompted.
FOR THE CLINICIANEnglish
English
Translation will appear here.
What the microphone heard
Clinician starts. Speak in short phrases, then let the translation finish. Watch the highlighted turn before speaking. The microphone pauses during translated playback, so speaking over it can lose words. Tap either speaker to correct a turn; turn off Hands-free turns for manual control.
One room microphone cannot reliably identify who is speaking. Automatic turns use quiet-time estimates, not speaker or language recognition. If there is no translation or the turn is wrong, use the speaker buttons. Keep this page in the foreground; leaving it ends the session.