Public audio devices
Compatible calling apps and browser calls select TCVoiceAI Mic and TCVoiceAI Speaker, the two user-facing routes for translated or original audio.
Virtual microphone translator
TCVoiceAI uses app-owned virtual audio devices for Translated voice, Original audio, and normal pass-through while the main app is idle. The public devices are TCVoiceAI Mic for outgoing call audio and TCVoiceAI Speaker for incoming partner audio.
One-sided setup: The background helper only relays audio between the selected physical and virtual devices. It does not recognize speech, create transcripts, translate, record, store, inspect, or upload conversation content, and it does not contact the server.
Signed and notarized for macOS. The production installer is signed with Developer ID and checked by Apple’s automated notarization service, allowing macOS Gatekeeper to verify its integrity and confirm that it is not known malware.
Setup steps
Choose an installed app in TCVoiceAI for exact device steps, or configure FaceTime automatically with permission.
The main app takes exclusive ownership of the audio routes only while Translated voice or Original audio is running.
Translated voice speaks translations. Original audio keeps real voices and optional History; bilingual sessions can switch modes live.
When the session stops or the app closes, the background helper resumes original-audio pass-through so the calling app can keep using the virtual devices.
Product details
Compatible calling apps and browser calls select TCVoiceAI Mic and TCVoiceAI Speaker, the two user-facing routes for translated or original audio.
During Translated voice, the foreground app recognizes and translates both directions, then sends Mac voice playback to the correct listener.
After you review a reply suggestion in Translated voice, Speak to partner sends its partner-language audio through the same TCVoiceAI Mic route.
During Original audio, pass-through remains live while optional History stores a bilingual translation or same-language transcript without spoken translation.
A bilingual session can change route ownership between Translated voice and Original audio without stopping. If the new route cannot complete, the previous mode is restored.
The foreground app gives each call direction its own recognition and translation flow. Translated voice also has a separate playback queue for each direction. Both directions can progress at the same time, while phrases remain ordered within their own direction.
While the app is Ready, Stopped, or closed, the helper relays the selected physical microphone to the virtual mic and virtual speaker audio to the selected physical output.
The helper handles audio buffers only to keep the route working. It performs no speech recognition, transcription, translation, recording, storage, content analysis, upload, or server request.
TCVoiceAI Mic Feed and TCVoiceAI Speaker Tap are internal endpoints used by the app and should not be selected in calling apps.
From the macOS app
Related guides
Start free with a monthly allowance and no payment method required.
Your conversations stay private by design.