Speech made simple

Frequently Asked Questions

My phone has live translation built in - why do I need VocalStack?

A phone runs a small speech model that has to fit on the device and the battery. That is fine for short, clear, everyday sentences. It is not fine for the things that actually go wrong in real speech: uncommon words and names, words that sound alike (“their” and “there”, “sail” and “sale”), technical terms, accents, people talking over each other and background noise. The model guesses, the translation is built on the guess, and no one in the room can see that it was wrong.

VocalStack runs on servers, so the models are full-size and every line is checked against the context after recognition. Upload a script or slides first and the names and terms in it win over what was misheard. If something still comes out wrong, you can fix the line while it is live and every listener sees the correction.

A phone also hears the room, not the speaker, and it translates for one person, so fifty phones give fifty different transcripts. VocalStack takes audio straight from the mixing desk, stream or microphone, transcribes the speaker once, and everyone reads or listens to the same thing in their own language. The translation is read aloud in the speaker’s own voice, so listeners hear the person on stage, not a generic robot voice.

Do I get charged per listener? How does pricing work?

No. Listeners are free, however many there are, in whatever language they read or listen. You pay for the audio that is transcribed, by the minute, and that is the whole bill: translation into up to 10 languages, spoken translations, broadcasting, summaries and exports are included.

See pricing and plans

Files and live transcriptions have different rates, and a monthly plan brings the rate down and includes credit. Every plan has the same features, and the API is charged at the same rate as the Dashboard.

What do listeners need to follow along?

A link or a QR code, and a phone. There is nothing to install and no account to create. A listener opens the link in a browser, picks a language, and reads the transcript as it is spoken. If they prefer to listen, each translated line is also read aloud. Change language at any time; the transcript so far is re-read in the new one.

A listener who joins late sees everything said before they arrived, then follows live. If the connection drops, the page catches up on its own when it comes back.

Does the audio have to come from a microphone?

No. A live transcription can hear a microphone on any device, an RTMP feed from OBS or a broadcast encoder, a YouTube live, an internet radio station or any URL that serves a stream. For a stream, VocalStack pulls the audio itself: nothing runs on your side.

One broadcast link can be fed by several transcriptions, so a conference with two stages, or a station with a radio stream and a YouTube live, still shares one link.

Can I control who sees the transcript?

Yes. A broadcast is public by default: anyone holding the link can follow it. Make it private and set a password, and only people you give both to get in. Rename the link later and the old one keeps working. You can also see who is listening, in which language, and when they joined.

Does a listener hear the real speaker or a synthetic voice?

Either. By default each speaker gets a distinct stock voice of matching gender, so a panel still sounds like a panel. With the speaker’s permission, you can clone their voice from thirty seconds of audio, and the translation is then read in their own voice in every language. A cloned voice also means the speaker is recognised and labelled by name in future recordings. You can assign a voice to each speaker, or mute one.

What is kept after the event?

Everything. The recording, its transcript with corrections and translations, and the audio are stored in your library. A title and summary are written once the words are final, in every language. You can play it back exactly as it arrived, search across every transcript in your language, and export it as SRT subtitles, PDF, Excel or JSON.

Can I build this into my own product?

Yes. Everything the Dashboard does is available through the API, at the same rate: realtime transcription over a WebSocket, live transcriptions with broadcasts, translation and spoken translations, file and link jobs, and an MCP server. Official SDKs for TypeScript and Python are included. Listeners can subscribe to a recording from your own app with a share code and no API key.