Live translation
Configure source audio, source language, and target languages for a translated live room.
Live translation is configured in the room setup at /translation/streamer. The Translation Streamer feature must be enabled, and the user needs the Streamer or member access shown by the dashboard. The room itself is created and started through the existing live-stream workflow.
Configure a translation
- Open /translation/streamer and create or edit a room.
- Open the Translation section in the advanced room settings.
- Enable translation.
- Select the source language.
- Select the source audio: the streamer’s audio or a configured ingress source.
- Select one or more target languages.
- Check the language count allowed by the organization’s plan.
- Save the room setup and start the broadcast.
The source language cannot also be selected as a target. At least one target language is required when translation is enabled. The server validates the language limit again when the translator starts, so a selection that exceeds the plan is not accepted by starting the room.
Supported languages
The current translator language list contains:
- Arabic
- English
- German
- French
- Spanish
- Turkish
- Italian
- Portuguese
- Russian
- Chinese
- Japanese
- Korean
- Dutch
- Polish
Arabic is the right-to-left language in the current language definition. The language names shown in the dashboard are localized according to the selected dashboard locale.
Choose the correct audio source
Use the streamer source when the speaker’s microphone is connected to the person running the room. Use an ingress source when the room receives audio from a paired hardware or professional source. Confirm that the selected source is producing audio before starting translation.
If the selected source disappears, refresh the ingress list and choose an available source. Starting translation without a valid source can result in no translated output even when the room itself is live.
During and after the broadcast
During a live room, monitor the source audio and the translated lanes separately. Translation is an additional processing path; a working original stream does not by itself prove that every target language is working.
After the session, use /translation/sessions to review the persisted translation session and its language outputs. Recordings, restreaming, and notifications are configured separately.
Common issues
- Translation is unavailable: check the organization feature and the user’s role.
- A target language cannot be selected: the target may be the source language or the plan cap may have been reached.
- No translated output: verify the selected source audio and that the source is speaking.
- The translator will not start: review the server validation message and remove targets above the plan limit.
- Arabic text direction looks unexpected: confirm that the target language is Arabic and test the public listener view.