BETA Now translating in 8 languages

Speak any language.
Hear it back in your own.

Duo Voice transcribes your speech and translates it in real time — directly in the browser. No installs, no setup, no waiting for a sentence to finish. Just open the page and talk.

Browser-only Simple sign-up ~200 ms latency
duo-voice.com / studio
Duo·Voice
Recording
Live
48 kHz
ENEnglish RURussian
● Original · EN MIC
14:02:08MICSo the way our pipeline works is we split incoming audio into small chunks.
14:02:14MICEach chunk gets transcribed and translated in parallel — there's almost no wait.
● Translation · RU LIVE
14:02:08МИКНаш конвейер работает так: мы разбиваем входящий звук на небольшие фрагменты.
14:02:14МИККаждый фрагмент распознаётся и переводится параллельно — почти без ожидания.
What it does

Three things, done in real time.

Built for moments when you need to understand the other person quickly — or make yourself understood — without juggling apps or waiting for a translation to load.

Voice → Text

The system transcribes your speech as you talk, with virtually no delay. Useful for dictation, notes, captions, and capturing thoughts on the fly. Stays accurate even with background noise — and you can copy, edit, or save instantly.

"Let's review the Q3 numbers…"

Text → Translation

Translation appears side-by-side with the source. Speak in one language, read in another — updates flow continuously, so the conversation stays natural and never has to pause.

ENCould we move the meeting to Thursday?
RUМожем перенести встречу на четверг?
ENSure, that works for me.

Flip mode

Built for live conversations. Swap the direction with one click — speak Russian and read English, then flip and reply the other way. Ideal for interviews, consults, and online meetings where two people need to understand each other.

ENEnglish RURussian
Dual-channel pipeline

Mic + system audio, in parallel — no interference.

Duo Voice can listen to your microphone and your system audio at the same time, transcribing both. So a video call becomes a two-way bilingual conversation — you talk, they talk, and both sides get understood in real time.

Microphone
→ stream A
System audio
→ stream B
Merged transcript
Two streams, one continuous record — speakers labeled, timestamps aligned.

Independent processing

Each channel is processed on its own pipeline, so the streams don't bleed into each other. Recognition stays clean even when you and the other person speak at the same time.

Speaker-aware transcripts

Every line in the transcript is tagged MIC or SYS — so a saved conversation reads like a script, not a transcript blob.

Built for video calls and lectures

Useful in webinars, interviews, and online classes — listen to the speaker, talk back, and get a clean record of both sides. No virtual cables, no extra software.

How it works

Five steps. About ten seconds.

01

Pick two languages

Choose your source and target — eight to start, more on the way.

02

Hit Record

Browser starts streaming audio. Mic and system audio toggle independently.

03

Speech becomes text

Audio is split into small chunks and transcribed as you go.

04

Translation appears

The translated text streams in beside the source — no waiting for full sentences.

05

Flip anytime

Swap the direction with one click. Save, copy, or edit the transcript.

Who it's for

Useful anywhere voice meets language.

If your work involves voice, text, or a second language, Duo Voice was built for the moments where the gap between them slows you down.

Talking with people abroad

Skip the awkward back-and-forth in two browser tabs.

Learning a language

See your speech in two languages at once — instant feedback.

Voice notes

Capture ideas verbally, get clean text back, ready to paste anywhere.

Online meetings

Two-way live captions during Zoom, Meet, or any browser call.

Subtitles & captions

Generate clean subtitles in any of the supported languages.

Educators & creators

Lecture, podcast, or stream — and walk away with a transcript.

Try Duo Voice while we're still in beta.

Free during the beta. No installs, simple sign-up — open the page, allow the mic, start translating. We're shipping new languages, punctuation, and model improvements every week.