How it works

One speaker in. Every language out.

AnyLang takes the audio you are already sending to the room, translates it once for each language somebody asked for, and delivers it to their browser as spoken audio, running captions, or both.

The core idea

Translated once per language, not once per listener

Most of what makes live translation expensive is doing it again for every person in the room. AnyLang opens one translation per language and everybody who wants that language joins it, so the cost follows the number of LANGUAGES you offer, not the number of people listening. A language nobody selected is never opened at all.

  • One paid translation per language
  • Listeners on a language cost nothing extra
  • A language with no listener is never opened
  • Speech and captions come from the same stream
Languages available
50+
Output languages per session
Up to 5
Maximum session length
3 hours
What a listener installs
Nothing

What is happening underneath

Four moving parts, and none of them is on the listener's phone.

Audio comes in

Either from a browser capturing a microphone or a mixer, or from an encoder such as OBS or vMix pushing a stream. Both end up in the same place.

It is translated per language

A translation opens for a language the moment its first listener arrives, and closes again shortly after the last one leaves. Music is recognized and left alone, so singing is not turned into invented lyrics.

Listeners join with a code

Six characters, or a QR code, or a permanent link that always points at whatever you are broadcasting now.

Captions ride along

The text is a by-product of a translation that is already running, so captions in a language somebody is listening to cost nothing extra.

Credits are spent by the minute

One credit is about a minute in one language. What moves it is how much is actually spoken, so quiet stretches and music cost less than solid speech. The balance updates while you broadcast, so buying more mid-service reaches the running session.

Nothing is stored on the listener

No app, no account, no sign-in. A listener is anonymous and is gated only by knowing the session code.

Four steps, start to finish

  1. 1

    Create an organization

    Sign up, say what kind of organization you are, and you are given 30 credits to try it with. The organization type is what tunes the translation for a sermon, a conference or a council meeting.

  2. 2

    Choose your languages

    Pick the language you speak and up to five to translate into. You can also offer captions in the spoken language, which is what makes the same setup useful to attendees who are hard of hearing.

  3. 3

    Start and share the code

    Starting gives you a six-character code and a link. Put either on a screen, in a program, or on a printed poster with a QR code.

  4. 4

    Speak

    That is the whole job. Listeners come and go on their own, translations open and close to follow them, and you are billed for the minutes that actually happened.

Common questions

How much delay is there?

Translated speech is held very briefly before it is played, so a whole clause arrives at once instead of arriving word by word. Captions in the language being spoken are not held back at all, because there is no translated audio for them to keep pace with.

What audio source gives the best result?

The cleanest feed you already have: a lapel or headset microphone through your mixer. A room microphone picks up echo, applause and air conditioning, and the translation follows whatever it is given.

Can I use my own terminology?

Yes. An organization keeps a glossary of names and terms, plus notes for the session about to run. Both are applied to the translations that open after you save them.

What happens if my browser crashes?

The broadcast pauses rather than ending, and listeners keep their connection. Reopening the console picks the same session back up. Only a deliberate Stop ends a session and releases the code.

Is there a free trial?

Every new organization is given 30 credits, which is 30 minutes of translation in one language. No card is needed to use them.

Try it on a real session

Two minutes to set up, 30 credits to spend, no card required.