Choose the audio source
Use a headset, room microphone or mixer. Where supported, capture computer audio, mix microphone and tab audio, or use a remote presenter. Panel capture can select the active speaker from separate microphones.
Simultaneous AI translation: One speaker.
Up to ten languages, live.
SalamAI translates one speaker into audio and captions in up to ten languages at the same time. Use it for conferences, presentations and training. Listeners follow on their phones or a shared screen.
Set up your first sessionIn short
What is simultaneous AI translation?
Updated
Simultaneous AI translation translates a speaker into other languages while they talk, without waiting for pauses. In SalamAI’s one-way mode, one audio source feeds the console and up to ten output languages run at the same time. Each language is its own channel with its own volume, captions and audience link, and each is billed separately. Listeners scan the QR code for their language and hear translated audio on their phone, or read captions on a shared screen. It suits keynotes, conferences, training and sermons, where one person speaks to many. It is not instant: audio on phones typically arrives about 6–20 seconds later, so test with your speakers before the event.
How much delay should you plan for?
Plan for translated audio on audience phones to arrive about 6–20 seconds after the speaker. The delay comes from recognising the speech, translating it and delivering the audio. It suits keynotes, sermons, lectures and presentations, where one person talks for a while. It is harder for quick back-and-forth, so during question time ask each person to finish their point, or have the moderator repeat questions from the stage. Test with your own speaker, language pair and venue network before the event, and tell listeners at the start that translated audio on phones runs about 6–20 seconds behind the stage.
Which audio source works best?
The cleaner the audio, the better the translation. A headset or a microphone close to the speaker works well for one presenter. For a stage with several microphones, send the mixer output into the console computer, using USB audio from the mixer or a USB audio interface; a headphone adapter cannot capture audio. Where your browser supports it, you can capture computer audio, mix a microphone with tab audio or use a remote presenter, and panel capture can pick the active speaker from separate microphones. Do a sound check with the real speaker, then listen to a translated channel on a phone before doors open.
Read nextSupported translation languages →Translation for places of worship →
For conferences and presentations
Translate one speaker for many listeners.
Your audio
Mic, tab, panel or remote speaker
Up to ten outputs
Independent language channels
Every listener
Room audio, caption screens or phones
Control each language independently
Set volume, mute, speaker device and captions for each language. Choose Voice and subtitles in 70+ languages, or another translation type your team has enabled. Some playback options also let you adjust speech speed.
Show captions on separate screens
Open a caption window for each language. Show it full-screen or use a simple text view. Adjust the text size and choose Words or Sentences display.
Share one link per language
Each language has a QR code and link. Choose audio, subtitles, joining approval and link expiry. See audience phone listening →
Need two-way translation?
Translate a conversation between two people.
Common questions
Event translation questions
Can SalamAI translate into several languages at the same time?
Yes. One SalamAI console can run up to ten output languages at once, within your organisation’s limits. Each output language is billed separately. Language and voice choices depend on the translation type. Check supported languages.
How does the audience hear the translation?
Give listeners a separate Audience Companion link for each language. They open the link in a browser and listen with headphones. You can also show captions on a shared screen. See how audience listening works.
Is simultaneous AI translation instant?
No. AI processing and audio delivery add a delay. Audience Companion audio typically adds about 6–20 seconds. Test your language pair, audio input and audience phones before the event.
Plan your event
Event coming up?
Let everyone understand.
Tell us your languages, crowd size and audio setup. We’ll show you SalamAI live and suggest the right setup for your event.
Already have an account? Open the console →