Other’s Audio: Translate What the Other Side Says
Overview
In Other and Both modes, Sokuji needs the other side’s audio: the voices of a meeting, a video or a live stream. The desktop app takes it from a specific application or from everything your computer plays; the browser extension takes it from the current tab. The original sound keeps playing as usual, and the translation is shown as text, or also read aloud to you with Translation I hear.
1Choose the source (desktop app)
In Settings, under Other’s audio, pick the application that plays the sound, for example Microsoft Teams, or System Audio (All Applications) for everything your computer plays. Picking the application keeps notification sounds and music from other programs out of the translation.

Tip
An application often appears in the list only once it is playing sound. Join the call or start the video first, then click the refresh icon.
2Change it from the main window
Click the selected mode at the bottom left. In Other mode it shows Other’s audio, which you can switch on or off or point at another source, and Translation I hear.

3Hear the translation
To have the other side’s translation read aloud to you as well, switch on Translation I hear in the Speech block and pick the device and channel it plays on. In the desktop app this needs one application as the source: with System Audio (All Applications) the row is greyed, because the translation would be captured and translated again. It cannot change while a session runs. See Spoken or Text Only.

4Browser extension
In the extension, Other’s audio is the tab you are in, for example your meeting. Switch it on; the tab’s own sound keeps playing on the Default playback device chosen in the Speech block when the session starts.
5macOS permission
On macOS, capturing all system audio needs the System Audio Recording Only permission. If no sound arrives, Sokuji says so and offers Open System Settings. Picking a specific application instead also works.
Tips
- On Linux, wear headphones in Both mode, so that your microphone does not pick up the other side from your speakers. On Windows and macOS, echo cancellation removes what your computer plays from your microphone.
- Other’s audio is translated as text. It is read aloud to you only when Translation I hear is on.
Frequently Asked Questions
The other side’s words are not translated. What should I check?
That Sokuji is in Other or Both mode, that Other’s audio is set to the right application or to System Audio (All Applications), and that the call is actually playing sound.
Why is Translation I hear greyed?
In the desktop app, it needs one application as Other’s audio source: pick the application instead of System Audio (All Applications). It is also greyed in Me mode, and fixed while a session runs.
Can I change the source during a session?
Yes, in the desktop app. If the switch fails, Sokuji keeps the previous source; stop and start the session to change it.
Troubleshooting
Still stuck? Send us feedback from the account menu, or write to support@kizuna.ai. You can also ask in GitHub Discussions.