How much does it cost to translate audio?

Translating a recording uses 2 credits a minute for the transcription, plus 1 credit for each 10 minutes translated. Dubbing is charged by the characters voiced, as text to speech, plus the background separation or their own voices when you use them. Every price is shown on the button before you start, and a failed run gives the credits back. Your sign-up credits translate about 28 minutes. See pricing for credit packs and plans.

Can I translate audio to English?

Yes. Upload the recording to the audio translator and choose English; the language spoken is detected for you. You get an English transcript next to the original, English subtitles, and an English dub if you want one.

Which languages does the audio translator work with?

Recordings in most widely spoken languages are understood, and dubs can be made in 36 languages, including English, Spanish, Mandarin, Cantonese, Japanese, Korean, Hindi and Arabic, each with native voices.

Does the dub keep the original speakers’ voices?

Each speaker gets a voice of their own. By default it is a native voice of the new language matched to how they sound; switch on “Keep their own voice” and each line is turned into the speaker’s own voice from the recording.

Will the dub match the timing of the original?

Yes. Every line is translated to fit the time it took, read a little faster where needed, and placed where it was said. Lines that would still run long are marked so you can say them shorter before dubbing.

Can I get only the translated text or subtitles?

Yes. The translation comes first: download a bilingual DOCX or TXT, or SRT and VTT subtitles, without dubbing anything.

Can I translate the audio of a video?

Yes. Drop in an MP4, MOV, MKV or WEBM file: the sound is taken out on your device and translated, and the picture stays with you. Put the dubbed track back under your video in your video editor.

Can translation apps translate an audio recording?

Most translation apps listen to speech as it happens, through the microphone. To translate a recording you already have, upload the file here and you get the whole text with timings, and a dub if you want one.

How long can a recording be?

Up to 3 hours to translate, and up to 60 minutes to dub. For a longer talk, dub it in parts.

Are my recordings private?

Yes. Your recordings, translations and dubs are visible only to you and stay in your library for 30 days.

Audio Translator with AI Dubbing

languages to dub into: 36
gets a voice of their own: Every speaker
per recording, dubs up to 60 min: 3 h
with every translation: SRT, VTT, DOCX
check it before you dub: Line by line
or a match from the library: Their own voices

How do you translate audio?

To translate audio, upload the recording, choose the language to translate it into and start. The audio translator transcribes every speaker, translates each line to fit the time it was said in and picks a voice for each speaker. Read the bilingual draft, fix any line, then download the subtitles or dub it into a new MP3.

How to translate audio in three steps

The translation arrives first, so you see every line before anything is dubbed.

Upload a recording or a video
Drop in an MP3, WAV, M4A or FLAC file, or a video such as MP4 or MOV. From a video only the sound is used, taken out in your browser, and the video stays on your device. Recordings run up to 3 hours.
Read the translation side by side
Every line shows what was said and its translation, with the speaker and the time. Fix a word, give a line to the right speaker, or rename the speakers. Download a bilingual transcript or subtitles right away if text is all you need.
Dub it and download
Pick a voice for each speaker, or keep their own, and dub. You get the dubbed MP3 and WAV, the voices on a track of their own (and the background too when you keep it), and subtitles timed to the dub. Change a line later and only that line is dubbed again.

What this audio translator does for you

A translation you can check before it is voiced, voices that fit the people in the recording, and timing that still matches the original.

Translate audio line by line, side by side

Each line of the recording sits next to its translation, with the speaker and the time it was said. Click a line to edit it, click its time to hear the original from there, and your edits save as you go. Names, brands and numbers are easy to check before anyone hears them.

Every speaker dubbed in a voice of their own

Speakers are told apart automatically, and each one is heard and given a native voice of the new language, chosen by sex and age, from 400+ voices. Swap any of them, use a voice you designed, or keep the speaker’s own voice from the recording.

AI dubbing that keeps the timing

Each translated line is sized to the time the original line took, read a little faster where it needs to be, and placed where the line was said. Lines that would run long are marked before you dub, so a two-minute interview still lasts two minutes and every answer follows its question.

Music and background sound kept

When music or room sound plays between the lines, it is separated from the voices and kept under the new ones, at its own level. The background and the dubbed voices also download on their own, ready for a video editor.

A bilingual transcript and subtitles with every audio translation

Every translation comes with a bilingual DOCX and TXT, and SRT and VTT subtitles timed to the recording, before you dub anything. After dubbing, the subtitles follow the dub, line for line.

Change a line, dub only that line

Edits after a dub cost only the lines you changed: the rest of the audio is reused, so fixing a name in an hour-long talk takes seconds and a few credits, not a new dub.

Audio translation samples

Each sample is one of our own recordings, translated and dubbed by this tool exactly as it came out. Switch between the original and the dub at the same moment, and read both sides line by line.

Audio translation guide: from recording to dub

Get a recording ready to translate

A clean recording makes a clean audio translation. Keep the microphone close, let people take turns, and keep music under the voices rather than over them. Overlapping voices are the hardest part to tell apart.

Videos work as they are: the studio takes out the sound on your device and translates that. For a long event, upload the whole file; recordings run up to 3 hours, and dubs up to 60 minutes.

Translate audio to English, or English into other languages

The language spoken is detected automatically, so a Spanish, Chinese or Hindi interview can be translated into English without telling the tool anything. Choose it yourself when the first minute is music, or when a speaker has a strong accent.

For dubbing, pick one of 36 languages, including Mandarin and Cantonese, each with native voices. When a market speaks one form of a language, pick voices that speak it and read a few lines aloud to check the words feel local.

Check a translation before you dub it

A few minutes of reading turns a good translation into one you can publish. Work in this order and the review goes quickly.

1Rename the speakers, so every line shows who says it.
2Read the lines marked as too long and say them shorter in your own words.
3Check names, products and numbers against the original, clicking a line’s time to hear it.
4Give any line the transcription heard from the wrong person to the right speaker.

Voices: a match for each speaker, or their own

Each speaker starts with a native voice of the new language that matches how they sound. Change any of them from the voice library, or use a voice you designed or cloned yourself.

Switch on “Keep their own voice” to keep how a speaker sounds: each line is read in the new language, then turned into that speaker’s voice from the recording.

Two speakers with similar voices are easier to follow with clearly different dub voices, one higher and one lower.

Subtitles or AI dubbing: which to use

Subtitles keep the original voices and suit viewers who read along, such as interviews and documentaries. They come with every audio translation, before you dub.

A dub suits listening: podcasts, lessons, voice messages and videos watched on the go. Many creators publish both, a dubbed track for listening and subtitles for search and accessibility.

Publishing a dubbed recording

Each dubbed MP3 and WAV carries a machine-readable note that it was made with AI, and most video and podcast platforms ask you to label AI voices when you upload. Add the subtitles too: they help people watching with the sound off.

Keep the original next to the dub, so you can answer a question about what was really said.