SeiriVoice Live · Live, in the languages you need
Your live feed, in another language, as it happens.
SeiriVoice Live listens to the signal, transcribes it, translates it and returns it captioned or dubbed in real time. For broadcasts, events and parliamentary sessions where accessibility cannot wait for post-production.
- 01
Accessible from the first second
Automatic captions on the stream and in the room, with no stenography and no captioning shifts.
- 02
One signal, several languages
Translate in real time and deliver each language as subtitles or as a dubbed audio track.
- 03
Fewer hands, same control
One operator supervises the channels from a dashboard. The machine does the repetitive part.
In the field · 4K HDR Summit
An international conference, captioned and translated live.
Medina Media Events used SeiriVoice Live at the 4K HDR Summit: automatic captions for everyone following the sessions by stream, and translation on screens for the audience in the auditorium.

What it does
A complete voice chain.
Each channel combines the engines you need: recognition, translation and synthesis.
Live transcription
Real-time speech-to-text with diarisation: it knows when the speaker changes.
Simultaneous translation
Translates text on the fly and keeps the context of the sentence, not word by word.
Subtitles and closed captions
As video with burnt-in subtitles, as closed captions, or in a full-screen prompter view.
Synthetic dubbing
Returns the signal with a dubbed audio track. Voice cloning keeps the original timbre.
Configurable channels
Each channel has its own input, engines and outputs. Create and monitor them from the browser.
Standard protocols
In over SRT or RTMP, the contribution protocols your chain already uses.
How it works
One signal in. As many out as you need.
- 01
Input
Send audio, or audio and video, over SRT or RTMP.
- 02
Recognition
The speech engine transcribes with voice activity and speaker detection.
- 03
Translation and voice
Text is translated and, if you want dubbing, synthesised in the target language.
- 04
Output
Prompter subtitles, video with closed captions, or dubbed video.
Demo
SeiriVoice Live in action.
Technical sheet
Built for a control room.
- Inputs
- SRT and RTMP; audio or audio + video
- Outputs
- Prompter subtitles, re-speaking, video with closed captions, dubbed video
- Engines
- Speech-to-text, translation and text-to-speech, combined per channel
- Monitoring
- Real-time web dashboard per channel: status, subtitles per language and signal health
- Tuning
- Voice activity detection, boosted custom terms and configurable subtitle delay
- Voices
- Your own voices from the training portal
- Deployment
- In the cloud or on your own GPU servers
FAQ
Frequently asked questions
How much delay does it add?
Subtitle delivery delay is configurable per channel; captioning adds less than translating and dubbing. We tune it with you, on your feed.
Can I use it for captions only?
Yes. Each channel switches on only the engines it needs: transcription, translation, dubbing, or all three.
Does it work for in-person events?
Yes. The prompter view shows subtitles full-screen in the room, alongside the output for streaming.
What about recorded content?
For files, transcription and translation are part of Seiri. SeiriVoice Live brings that same intelligence to the live signal.
Next step
Send us your feed. We’ll tune the delay with you.
A trial on your own live signal, with the languages and outputs you need.