Everything Tolka Live can do — live AI translation, captions, voice, integrations, pricing and the practical details. Something missing? Email [email protected].
Tolka Live is real-time AI speech translation for live events. It listens to the person speaking and shows the translation as live captions — on a projector, inside ProPresenter, in an OBS livestream, or on the audience's own phones — and can also speak the translation aloud with a natural AI voice. It supports 44 target languages and is used at church services, conferences, lectures, theatre and broadcasts.
You feed it a microphone signal (the built-in mic works, a feed from your mixer works better). The audio is streamed to a cloud AI that translates speech directly, and the translated text appears word by word while the person is still talking — typically within a second. There is nothing to script or prepare: the source language is detected automatically.
Three ways to run it: a macOS app (Apple Silicon and Intel), a Windows app (Windows 10/11), and a web app that runs directly in the browser at tolkalive.se/app — nothing to install. The audience needs nothing but a phone browser.
No. Create an account, get a free 5-minute trial (no credit card), open the web app in your browser, press Start and speak. The desktop apps are available on the download page when you want them.
Any live setting with a multilingual audience: churches and international congregations, conferences and corporate events, lectures and courses, theatre, municipal meetings, and livestreams. If people in the room don't all share a language, Tolka Live gives each of them the words in their own.
44 target languages: Swedish, English, Icelandic, Norwegian, Danish, Finnish, German, Dutch, French, Spanish, Portuguese, Italian, Polish, Romanian, Hungarian, Czech, Greek, Russian, Ukrainian, Turkish, Arabic, Hebrew, Hindi, Bengali, Tamil, Thai, Vietnamese, Indonesian, Filipino, Chinese, Japanese, Korean, Swahili, Amharic, Yoruba, Igbo, Hausa, Zulu, Xhosa, Shona, Sesotho, Somali, Lingala and Afrikaans. The source language is recognized automatically.
Yes. The web app can run up to five languages in parallel from one microphone. Each language gets its own permanent link and QR code, so the Spanish-speaking part of the audience scans one code and the Ukrainian-speaking part another. Plan hours are counted per language (two languages use hours twice as fast).
Very good, but not perfect — like all AI translation. Individual words, names or numbers can come out wrong, especially with unclear audio or unusual expressions. Treat the captions as support for the audience rather than certified interpretation, and use professional interpreters where legally or medically critical accuracy is required. Clear speech into a good microphone gives the best results.
Yes. Every event can have its own instructions to the AI — for example a liturgical register for a service, or your industry's terminology for a conference. A built-in per-language profanity filter keeps captions clean by default (religious words are never censored), and you can add your own blocked words.
Captions start appearing while the sentence is still being spoken — word by word, typically with about a second of delay. The optional AI voice follows a moment behind the captions.
Yes, two ways. The recommended way is a web overlay in a ProPresenter Prop: you paste your personal caption link into a web object and the captions float over your slides — the operator can clear the Prop and it stays cleared, and captions persist across slide changes. Alternatively, captions can be pushed into ProPresenter's native Messages (with your own theme) and to the stage display via the network API. Setup takes minutes; see the step-by-step guide.
Yes. Add your caption link as a Browser Source and a transparent, styled caption overlay appears in the stream. Font size, colour, backdrop and position are adjustable with URL parameters.
The app has its own fullscreen caption view — connect a projector or TV and you're done. Any device with a browser (a smart TV, a spare laptop) can also show the caption page directly.
Yes. Show the QR code (the app can project it fullscreen, copy it as an image, or print it) — the audience scans it and reads the live captions on their own phones, with adjustable text size and a scroll of recent lines. The link and QR code are permanent for your account, so a printed sign keeps working at every future event. Readers and listeners are unlimited and free.
Yes. When the host enables voice broadcasting, a Listen button appears on the phone page — tap it and you hear the translation spoken in a natural AI voice through your own headphones, like a personal interpreter channel.
The phone page shows every language currently being translated — they just tap theirs and the page switches instantly.
Yes. The caption links are hosted (tolkalive.se), so a stream team in another city can drop the same overlay into their OBS — no shared network needed.
Every plan includes every feature — the only difference is translation hours per month. Basic: €55/month for 15 hours. Standard: €90/month for 30 hours. Pro: €220/month for 75 hours. (Swedish pricing: 595/995/2 395 kr.) Need more in a single month? Extra hours (10 h for €33 / 349 kr) roll over until used. Prices exclude VAT.
Yes — every new account gets 5 minutes of free translation, no credit card required (one trial per computer). That's enough to test your microphone, see the captions on a phone and try ProPresenter.
Only active translation time. Pressing Mute (during songs, breaks, music) is free. If you translate into several languages at once, each language counts its own time — one hour of speech translated into two languages uses two plan hours.
No. Any number of people can read the captions or listen to the voice — on phones, projectors or in the stream — at no extra cost.
A computer (Mac, Windows or just a browser) and a microphone signal. The computer's built-in mic works for testing; for live events a feed from your mixing desk or an audio interface (e.g. Focusrite Scarlett) gives much better results. Mic gain and a noise gate are adjustable in the app.
Yes — translation runs on cloud AI, so a stable connection is required. Normal broadband or good 4G is more than enough; the audio stream is small.
Captions: effectively unlimited — hundreds of phones and screens are fine. Spoken voice on phones streams more data, so on a typical venue Wi-Fi keep voice listeners to a few dozen; captions scale further.
No. The speech is streamed to the translation AI, used to produce the translation in real time, and is not recorded or stored by Tolka Live. The website uses consent-based cookies (EU Consent Mode v2).
No — the AI runs in the cloud. If your connection drops briefly, the app reconnects automatically and continues.
It doesn't replace certified interpreters where they're legally required — but for the everyday multilingual audience it's instantly available, covers 44 languages at once, needs no booths or headset rental, and costs a fraction of professional interpretation. Many venues use both: interpreters for the critical parts, Tolka Live for everything else.
The guide covers setup step by step (installation, audio, ProPresenter, OBS, troubleshooting). For anything else, email [email protected] — real humans answer.
Tolka Live