How to translate your first voice message or audio file
Open Audio Translation and add a file, recording or link: recognition and translation start on their own. Tap play to hear the translation in a voice.
Updated November 2, 20264 min read
The example below is a voice message in a language you don't speak. You need tokens on the balance; prices are at the end of the article.
Step 1. Open Audio Translation and add the audio
On the Tools tab, tap Audio Translation. The screen is called Translate audio.
Add the audio in one of four ways:
Choose: an audio or video file from the device, up to 100 MB. From a video, the app takes only the sound.
Record audio: record right in the app and tap Done when finished.
Paste link: a link to YouTube, TikTok, Instagram or a direct file link, up to 2 GB.
Import from messenger: shows how to send a voice message from Telegram, WhatsApp, Viber or iMessage through the share menu of the iOS or Android app. In the share menu, pick Translate Audio in VocaLingo.
Formats and limits for every tool are collected in supported files and links. Sharing from messengers step by step: how to share audio from messengers.
Step 2. Wait for recognition and translation
Audio Translation starts on its own as soon as the file is added: there is no Translate button. The progress bar shows "Recognizing...", then "Translating...". The source language is detected automatically, you can't and don't need to choose it.
The result appears on three tabs:
Source text: what was said, and the detected language in the Language row.
Translation: the translated text and the voiceover player.
Summary: Brief, Key points and Chapters. Useful for long recordings.
Step 3. Check the target language
The target language is in the Translation row at the top of the Translation tab. By default it's the language you chose last time, or the app language on the first run. Tap the row, find the language with search, and the text is translated again automatically. Which languages are available: supported languages.
Step 4. Listen to the translation
The voiceover isn't created automatically: tap the play button on the Translation tab.
If you haven't picked a voice in the settings, for Russian, English, Chinese, German, French, Spanish, Italian, Japanese, Korean and Portuguese the app clones the speaker's voice itself if the recording is long enough.
For other languages it asks "Clone voice?" and shows the price in tokens. Clone voice makes the translation sound like the speaker, Use AI voice reads it with a standard voice.
Voices and providers can be chosen in advance with the settings icon at the top of the screen, see Translate Audio settings. If the voice sounds robotic, too fast or not like the speaker, see why a translation sounds robotic.
Step 5. Copy, share or reply
Under the translation you can copy or share the text, and the audio is shared from the player. There is no TXT, SRT or PDF export in Audio Translation. The What next? block offers:
Record a reply: starts a recording and translates your answer into the language of the original message;
Discuss in chat: opens the translation in AI chat;
Save voice: after a MiniMax clone, adds that voice to My Voices.
The result is saved to the History of the tool and shows up in Recent on the Tools tab. Without Premium, each tool keeps the last 3 entries.
How much the first translation costs
Recognition and translation cost less than one token per minute of speech; small amounts add up and are charged as whole tokens. The voiceover is charged separately when you tap play:
Voiceover | Price |
|---|---|
Standard ElevenLabs voice | 10 tokens per 1000 characters of translation, at least 1 |
Speaker's voice clone, automatic (10 languages listed above) | 20 tokens per 1000 characters, at least 5 |
Speaker's voice clone with MiniMax | 180 tokens for cloning plus 12 tokens per 1000 characters |
A voice saved to My Voices with MiniMax is not charged 180 tokens again. All prices are in token pricing.
If something goes wrong
"Please choose a file no larger than 100 MB.": the file is too large. Upload it to a cloud drive or a video site and use Paste link, links work up to 2 GB.
"Recording is too short. Please record at least 1 second.": record again and speak a bit longer.
"Processing failed. Please try again.": choose the file again. If it repeats, write to support and attach the file.
"Provider does not support the selected language, switched to ElevenLabs": the chosen clone provider can't speak the target language, so a standard voice was used.
"Not enough tokens to finish this action": top up the balance, see how to buy tokens.
Frequently asked questions
Can I fix a mistake in the translation before the voiceover?
Yes. Tap Edit on the Translation tab, change the text and tap Save, then play the audio again so the voiceover uses the new text.
The app recognized the wrong language. What can I do?
The source language can't be set by hand. Check the Language row on the Source text tab. If the recording is very short or noisy, try a longer and cleaner fragment.
Contact support
Describe the problem and we will reply by email. Add the file name or job ID if you have one.