One mic in.
Four languages out.

It takes one speaker from an input of your audio interface, translates the speech in realtime and speaks it again in up to four languages. Each language comes out on its own output channel, so it can go straight into a mixer, a stream encoder or an interpretation system.

For anyone who runs audio for multilingual audiences: general meetings and assemblies, conferences and live events, livestreams, internal town halls and hybrid meetings. The aim is simple: no one in the audience is left behind.

Up to four output channels, each in its own language, from one input. Channels 1 and 2 are on by default; channels 3 and 4 start switched off so nothing runs or bills by accident. Optionally, Speech Connect handles up to 16 translation languages.

No. Translation trails the speaker by a few seconds, sentence by sentence. For negotiations and other high‑stakes settings, book interpreters. It fits where a clean, automatic language channel is enough, for example on a livestream, at a general meeting or at an internal event.

Typically a few seconds, because it works sentence by sentence. A per‑channel limit, the max translation backlog, drops queued speech if a channel falls too far behind, so it catches up instead of drifting.

Each channel speaks with a female, male or speaker‑matched voice. You choose it per channel.

Yes, if you want. In Ducking mode the original stays underneath and is lowered while the translation speaks, like a TV voice‑over. In Isolated mode only the translation is heard, like an interpreter feed. You choose per channel.

Then that channel gets the original voice, not a machine re‑read of it. With a fixed input language, that channel doesn’t even open a translation session.

With auto‑detect, the language is detected as the speaker goes. A channel in the language being spoken gets the original voice. Detection needs the first words; a pass‑through delay on the original covers that gap, or you set the input language to fixed.

The output languages are those the realtime translation API offers for speech. Name your languages in the form and we’ll confirm them before the demo.

Each language can go to any physical output of your audio interface. From there it is an ordinary audio feed for your mixer, stream encoder or interpretation system.

Language, voice, output channel, Isolated or Ducking, duck depth, attack, hold and release, translation and original levels, a pass‑through delay for the original, the max translation backlog and a channel master.

Yes. A live transcript shows the original and every channel as it runs.

Yes. Ducking runs on the audio clock instead of a UI timer, so it stays accurate while the window is minimized or behind other windows.

A Mac with macOS 13.5 or newer (Intel or Apple Silicon), Chrome, Edge or Brave installed, an audio interface, an internet connection and a Speech Connect subscription. The app carries its own runtime, so nothing else needs installing.

A Windows version is in development. Today Speech Connect for the live desk runs on the Mac; tell us in the form if you are waiting for Windows.

Some multichannel interfaces expose only their first two inputs to the browser. Route the speaker to input 1 or 2 in the interface’s mixer, or use a macOS aggregate device.

The audio is sent to the realtime translation API for transcription, translation and speech. The API‑key is stored locally on the Mac, readable only by your user account, and the app only accepts connections from the computer itself.

Yes. Speech Connect is developed in Austria, in the European Union, and is GDPR‑compliant. The app stores its key locally and only accepts connections from the computer itself; this website sets no cookies and uses no tracking.

Speech Connect is a subscription service (SaaS). Each translated channel runs its own realtime session and is billed as its own session: four languages use four times the minutes. A channel whose language matches a fixed input language opens no session at all.

Speech Connect is offered as a subscription. The price depends on your setup and the number of languages. Choose “Contact sales” in the form and we’ll send you an offer.

The Speech Connect audience platform adds audio streaming to listeners, live subtitles, chat, voting and games. Ask us about it in the form.