More people watch your service online than you might think, and a good number of them do not speak the language you preach in. Adding live subtitles turns a one-language broadcast into something a much wider audience can follow — and it is considerably simpler than it sounds.
Why subtitles matter more online than in the room
On a stream you lose every advantage a building gives you. There is no printed handout, no neighbour to lean over and explain a word, no interpreter, and nobody to ask afterwards. There is also nobody to notice that a viewer has given up.
Text on screen is the one thing that travels. It also serves several audiences at once: viewers who are deaf or hard of hearing, people watching with the sound off in a shared house, anyone whose second language reads more easily than it listens, and — a large group nobody counts — people watching on a phone in a noisy place.
The two ways to do it
Broadly there are two approaches, and the difference matters.
Burned into the video. Captions are rendered into the picture itself, so every viewer sees the same text in the same language. Simple, works everywhere, and fundamentally single-language. Adding a second language means a second video feed.
Sent to each viewer's own device. The viewer opens a link alongside the stream and reads in whichever language they picked. One broadcast, many languages, no extra feeds — and each person controls their own text size, which matters more than it sounds on a phone.
The second is what makes a stream genuinely multilingual. It also degrades gracefully: if a viewer's phone is busy, the video keeps playing.
What you need
- A device to run it from — usually the same laptop you already stream with, or a second one if the streaming machine is busy.
- A clean capture of the speaker's audio. A feed from your mixer gives far better results than a room microphone, and this matters more than anything else on the list.
- A translation layer that turns speech into text and pushes it to viewers in real time.
Notice what is not on that list: no captioning agency, no interpreter booth, no specialist hardware, and no change to how you stream.
Where to put the link
This is where most churches lose their online audience, and it is entirely avoidable.
- In the stream description, as the first line rather than the fourth paragraph.
- Pinned in the chat at the start, and again twenty minutes in for people who arrived late.
- On a slide at the beginning, and as a small permanent corner graphic if your setup allows it.
- On the page the stream is embedded in, if you host it on your own site.
The failure mode online is identical to the one in the building: the invitation appears once, early, and everyone who joins later never learns it exists.
A note on latency
Live captions arrive a second or two behind the spoken word. That is normal, and for a sermon it is perfectly comfortable to read.
What matters more is that the stream itself is usually delayed too — often by considerably more than the captions. If your video runs several seconds behind reality and your captions run two seconds behind, the captions will arrive ahead of the speech, which is disconcerting rather than helpful.
Worth testing before you announce anything: open your own stream on a phone next to the captions and see which arrives first. If the captions run ahead, most streaming setups let you add a small delay to bring them into line. It is a five-minute job that transforms the experience.
Do not chase zero delay
Once the two are roughly aligned, stop optimising. Half a second of latency is worth nothing next to clean audio and correctly spelled names, and effort spent on the former is effort taken from the latter.
Feeding the system the proper spelling of the people and places in your tradition does more for a viewer's experience than any amount of tuning — and unlike latency, a mangled name is something they will notice every single week.
The archive is the larger prize
One thing worth knowing before you set this up. The live audience for most church streams is modest; the audience for the text afterwards, over months, is usually much larger.
A stream is watched once by whoever was free at that hour. A transcript, in several languages, on its own permanent page, is found by people searching a question months later — and search engines read text, not video. If you are going to the trouble of captioning live, keeping the result is the step that makes it worth having done.
Who is actually watching
Worth thinking about before you decide what to offer, because the online audience is not a smaller copy of the room. It is four different groups with four different needs.
- Housebound and unwell members. Often the largest group and the most loyal. They know the church, they know the preacher, and what they need most is reliability — the stream starting on time, every week, without them having to hunt for a link.
- Members who have moved away, or are abroad for a season. Frequently watching in a different time zone, which means most of them are watching the recording rather than the live stream. For them the transcript matters more than the captions.
- People considering visiting. They are assessing you, quietly, and they will not tell you they were there. If they cannot follow the language, they will not come.
- People who found you by searching. They have no connection to your church and arrived because of one question. They are the group most likely to need another language, and the group least likely to ask for it.
The first two groups need consistency. The last two need to be able to read in their own language without asking anybody for anything — which is precisely what device-side subtitles and a kept transcript provide.
Try it at your next service
InMyTongue streams live subtitles and spoken translation to every viewer's own device, in the language they choose, straight from a browser with nothing to install — and keeps the transcript in every language afterwards. New accounts include free credit and no card is needed to try it. See how it works.