The congregation is already multilingual. Three ways to make Sunday work without exhausting your one bilingual volunteer — and the audio decisions that make or break the captions.
Most multilingual churches run on a miracle of volunteerism: one gifted bilingual member in a back row with a headset mic, simultaneously interpreting the sermon week after week. Simultaneous interpreting is a professional skill — doing it every Sunday, for free, in one direction only, is how churches burn out their most valuable volunteer and quietly cap which languages they can serve. Live translation stops being a staffing problem the moment the sermon becomes text — and the setups that get you there cost less than one Sunday's coffee budget.
The pattern repeats across denominations and countries. A congregation grows a second-language community — Spanish, Portuguese, Korean, Ukrainian, Swahili. Someone bilingual steps up. For a while it works. Then the interpreter travels, gets sick, or simply tires; the third language arrives and there is no volunteer for it at all; and the newest families sit through the sermon nodding at the rhythm of a language they only half hold.
The bottleneck was never willingness. It is that live interpreting is scarce, one-language-at-a-time labor — while a live transcript, once it exists, translates into many languages at once for close to nothing.
A laptop at the soundboard runs Unicaption and sends captions to a screen or the lower third of the existing projection. Best when one language covers most of the need — English preached, Spanish on screen. Zero effort from the congregation, one shared target language.
Each person who wants translation opens Unicaption in their phone's browser — no install — and picks their own language from 100+ options. This is how one sermon reaches five languages at once. The phone listens through its microphone, so it works best with a clear PA and seating reasonably near a speaker; test from the actual pews before promising it.
Your interpreter keeps interpreting — into headsets, or for the language that deserves a human voice — with the live transcript and translation on a screen in front of them. Scripture references, names, and the sentence they half-missed are on screen instead of gone. The role becomes sustainable instead of heroic.
Churches mix these: a screen for the biggest second language, phones for the long tail, the volunteer for the community the church most wants to welcome personally.
Every disappointing caption demo in a church traces back to the same cause: the tool was fed the room instead of the voice. Worship spaces are close to a worst-case acoustic environment —
The fix is to bypass the room entirely. Take a feed from the soundboard into the captioning laptop — the preacher's mic channel, not the full music mix — or, if the service is livestreamed, open your own stream and caption its tab audio: the encoder already has the clean mix you spent money on. (That same trick powers captioning live streams generally.) Only fall back to a room microphone when neither exists, and then keep it close to the pulpit. More squeezing-out-accuracy technique lives in our guide to improving live caption accuracy.
Sermons are dense with words no general-purpose engine should be expected to guess: scripture names, book-and-verse references, denominational vocabulary, and the proper names of your own congregation. These are also the words a congregation notices instantly when they come out wrong.
Church tech is run by volunteers on donated hardware with no Tuesday rehearsal time. The tool has to respect that:
The free tier gives you 30 minutes every week with no credit card — one sermon's worth. Plans start at $9/month, which is a different conversation with a church board than interpreting equipment ever was. Setup takes minutes; the walkthrough is in Getting Started.
| Your situation | Start with | What you need |
|---|---|---|
| One large second-language community | Projection screen | Laptop at the board, a soundboard feed, screen space |
| Many languages, scattered need | Congregants' phones | Good PA, a tested browser link, one announcement from the front |
| A faithful interpreter you want to keep | Interpreter-assisted | A screen in their line of sight, the dictionary preloaded |
| Livestreamed services | Caption the stream's tab audio | The cleanest signal you own — already mixed |
Run a live translation tool from a clean audio feed. Unicaption takes the preacher's audio — ideally a soundboard feed or the livestream's tab audio — and shows the transcription with live translation side by side in any of 60+ languages. Put it on a projection screen for one shared language, or let congregants open it in their phone browsers and each choose their own.
Yes. Unicaption runs in a phone browser with no app install, and each person picks their own target language — so one sermon can be read in five languages at once across the room. It works best with a clear PA system and phones seated within good earshot of a speaker; test from the actual pews before announcing it.
Audio, almost every time. Sanctuaries add reverb, music under speech, and distance between the voice and the microphone — all of which degrade any transcription engine. Fix it by feeding the captioning tool the preacher's channel from the soundboard, or your livestream's audio, instead of a room microphone. With a clean feed, sermon captions perform like meeting captions.
It shouldn't — it should relieve them. Machine translation covers the languages you have no interpreter for and the weeks your interpreter is away, while the live transcript makes their own work easier: scripture references and names on screen, nothing lost to a missed half-sentence. Many churches keep the human voice for their core second-language community and use captions for everything else.
Live sermon translation from a clean audio feed — 60+ languages, browser-based, nothing stored. 30 free minutes every week.
Start Free Trial →