Skip to content
English
Submit a request
  • There are no suggestions because the search field is empty.

Translating audio with Lara (audio-to-audio)

View pricing table

Lara supports audio-to-audio translation through a dedicated API and web interface, allowing you to translate spoken content directly from one language to another and receive the result as an audio file.

This feature is designed for programmatic and voice-based use cases where both the input and the output are audio.

 

What is audio-to-audio translation

With audio-to-audio translation:

  • the input is an audio file
  • the output is a translated audio file
  • speech is translated end-to-end without manual intermediate steps

You do not need to separately manage speech-to-text or text-to-speech processes.

 

Access and availability

Audio-to-audio translation is available through:

  • Lara's web interface
  • Lara's SDKs

 

Supported languages

Audio-to-audio translation supports a wide range of languages and locale variants.

Because language availability depends on the underlying speech capabilities and may evolve over time, the most up-to-date list of supported languages is maintained in the developer documentation.

For the current list of supported languages and locale codes, refer to:
https://developers.laratranslate.com/docs/translate-audio

 

Supported audio formats

Audio-to-audio translation accepts the following input formats:

  • .wav
  • .mp3
  • .opus
  • .ogg
  • .webm

Please ensure that uploaded files are provided in one of these supported formats to avoid processing errors.

 

File requirements

Uploaded audio files must meet the following requirements:

  • Maximum file size: 200 MB
  • Maximum duration: 2 hours

Files exceeding these limits will not be processed.

 

Typical use cases

Audio-to-audio translation is commonly used for:

  • voice-based applications and assistants
  • audio content localization
  • speech-driven workflows
  • multimedia and accessibility scenarios

 

Voice Cloning

Audio-to-audio translation now supports Voice Cloning. Instead of using a predefined target voice, Lara analyzes the acoustic profile of the source speaker — including timbre, pitch, and vocal nuance — directly from the input audio, to generate target-language audio that closely mimics the original speaker's voice. No separate voice sample is required.

Voice Cloning is currently available through Lara's SDKs.

This is useful when continuity of speaker identity matters, such as:

  • multilingual media and video localization
  • training and onboarding content
  • presentations and voiceovers
  • other voice-driven workflows where the audience should recognize the same "voice" across languages

Usage and pricing

Audio-to-audio translation usage is calculated based on audio duration.

Usage is billed per minute of processed audio.

Current rates:

Audio-to-audio translation (standard voice) — € 0.08 / $ 0.08 per minute
Audio-to-audio translation with Voice Cloning (SDK only) — € 0.40 / $ 0.40 per minute

Examples:
1 minute of audio (standard) → €0.08 / $0.08
60 minutes of audio (standard) → €4.80 / $4.80
1 minute of audio (Voice Cloning) → €0.40 / $0.40
60 minutes of audio (Voice Cloning) → €24.00 / $24.00

 
Developer documentation

Detailed technical information — including supported audio formats, request parameters, limits, SDK usage examples, and supported languages — is available in Lara's developer documentation.

For the most up-to-date implementation details, refer to:
https://developers.laratranslate.com/docs/introduction

 

 

This article is about:

  • What audio-to-audio translation is in Lara
  • How to access it via Lara's web interface, API, and SDKs
  • Voice Cloning support and where it's available
  • Supported audio formats and file requirements
  • Where to find the supported languages
  • How usage is calculated and billed