Coverage
How much was available
The share of the service for which the translation pipeline was actually up.
Live AI translation for continuous speech
One cable, and the room hears it.
Yazyk translates the speech feed your sound desk already produces and sends it live to the FM receivers the room already owns, to phones, or to captions, so people hear it in their own language. Afterward it reports how much was actually covered.
Services, conferences, classrooms, courtrooms: one speaker, an audience that cannot stop and ask, and a delay short enough to follow. Churches are where Yazyk runs today, and the first place it had to be right.
Coverage, delay, and interruptions are measured from the run log today. Paid billing is not live and no card can be charged; the figures on the pricing page are what we intend to launch with.
A checkable number about your own last event.
Not a claim about quality, which no committee can judge. Coverage, listener delay, and interruptions are measured from the run log, so what went wrong and when is on the record instead of an event clock standing in for a result.
Experience
What listeners heard
The measured listener delay, and how much translated speech was produced.
Interruptions
What stopped, and when
Every detected outage, with its clock time and untranslated duration.
Receivers, phones, and captions.
They serve different people in the same building. Yazyk supports all three deliberately; no single delivery path is the product's position.
FM receivers
Use the system already in the room
A cable takes translated audio from the computer to the room's existing FM transmitter and bodypack receivers.
Phones
Listen in a browser
A phone gives listeners another audio path when that fits the event and the network available in the building.
Captions
Read translated or source speech
Translated captions are another delivery format. Same-language captions are accessibility, not translation, for a different need in the same room.