Live Sessions
- One-way sessions translate and caption one speaker live
- Two-way conversations support hands-free voice detection or push-to-talk
- Remote Speaker lets a presenter use a temporary browser link from their phone
Live Speech Translation
Say It Once. Everyone Hears It in Their Own Language.
One person speaks. SalamAI translates the speech live, plays it aloud, and shows subtitles. It runs in the browser, so the speaker and audience do not need to install an app.
Talk to an Expert
SalamAISolution overview
Live Translation for Multilingual Rooms
SalamAI supports up to 100 languages through its broader translation engine, with 13 curated languages offering spoken output today. A single live session can deliver up to five output languages.
Session modes
Use the Setup That Matches the Room
Audience and Follow-Up
- Companion Mode provides translated audio and synchronised subtitles on audience phones
- Encrypted session history
- Optional summaries and recordings
- Text-only caption mode for OBS
Where it helps
For Any Room Where People Do Not Share One Language
Built around the multilingual reality of Malaysia and Asia Pacific.
Conferences and town hallsClinics and public countersWorship and educationTours and customer support
How it works
Three Steps. Nothing to Install.
1. Choose a microphone, mixer, browser tab, app audio, or remote speaker2. Select the output languages, volume, and audio device3. Start speaking to show subtitles and play translated speech4. Share Companion Mode when the audience needs phone listening
Security and honest limits
Clear Controls, With the Trade-Offs Stated Plainly
Security
- Two-step sign-in and invite-only teams
- Database-enforced tenant isolation
- Tamper-evident audit logs
- Browser encryption before transcript and recording uploads
- Local search and playback decryption
- Short-lived, single-purpose provider credentials
Current Limits and Billing
- Companion Mode currently supports one-way sessions
- Companion Mode has roughly 6 to 20 seconds of delivery delay
- Remote Speaker credentials are temporary and single-use
- Billing is based on active translated audio minutes
- Seats, listeners, and silent wall-clock time are not billed
