Online audio translation relies on two distinct technical steps: automatic transcription (speech-to-text) followed by the translation of the resulting text. The quality of the outcome depends almost entirely on the speech recognition model used beforehand. We have selected seven free tools by evaluating them on transcription accuracy, the number of target languages, file audio privacy management, and the actual limitations of the free tier.
1. Google Translate (audio mode)
Further reading : How to Easily Perform Unit Conversions
Google Translate remains the most accessible option for quick audio translation. The microphone mode allows for real-time voice input, and copying and pasting a transcription obtained elsewhere provides access to translation in over 130 languages. For those looking for free solutions to translate audio, this is often the first reflex.
The main limitation: Google Translate does not directly accept an audio file as input. You must use voice dictation or transcribe separately. The quality of translation varies greatly depending on the language pairs, with significantly better results on English-Spanish or English-French pairs than on Chinese to French.
See also : How to Easily Convert Milliliters to Centiliters: A Practical Tip for Gluten-Free Foodies
2. ElevenLabs Dubbing (free tier)
ElevenLabs offers an audio translation tool that preserves the tone of the original voice. The free tier accepts .mp3, .wav, and .m4a files, with a limit of one minute per file. The preservation of vocal tone in the target language sets this tool apart from most competitors, which replace the voice with a generic synthesis.
The announced support exceeds 90 languages. However, the one-minute constraint makes the tool unsuitable for long recordings (meetings, conferences). For short excerpts, promotional podcasts, or demo clips, the result is usable without manual retouching in most cases.
3. Kapwing Audio Translator
Kapwing combines transcription, translation, and subtitling in a single workflow. The tool automatically generates a transcription, translates it, and then offers to integrate subtitles into the source video. The free tier covers over 40 languages and includes synthetic voice options for dubbing.
The search and replace functions in the transcription allow for correcting proper names or technical terms before translation, significantly improving the final accuracy. The free limitation concerns export duration and the watermark added to videos.
4. Whisper (OpenAI) via no-code tools
The Whisper model from OpenAI remains the reference in multilingual transcription. Several no-code platforms (Make, Zapier, Pipedream) offer ready-to-use scenarios that chain Whisper for transcription and a translation model for text output. The free tiers of these platforms are sufficient for light use: a few meetings per month or short podcast episodes.
The technical advantage is modularity. We recommend this approach to users who want to control each step of the pipeline: choice of transcription model, translation engine, output format. The initial setup cost is higher than a turnkey tool, but the accuracy on languages like Chinese or Arabic is superior to most proprietary web solutions.
5. Notta
Notta targets meeting transcription and offers integrated translation. The application manages audio file imports, real-time transcription, and translation of the resulting text. The free tier provides a monthly quota of transcription minutes.
The native integration with video conferencing tools (automatic recording from Google Meet, Zoom) makes it a relevant choice for international teams. The translation occurs after transcription, allowing for proofreading and correcting the source text before launching the linguistic conversion.
6. VEED.io (online audio translation)
VEED.io works entirely in the browser. The tool accepts an audio file, transcribes it, and then translates the result into the desired language. The free tier imposes a watermark and a duration limitation, but the transcription-translation-subtitling chain remains functional without paid registration.
A technical point to note: VEED.io allows exporting the translated transcription in SRT or VTT format, which facilitates integration into video post-production workflows. For content creators publishing in multiple markets, this compatibility with standard subtitle formats makes a difference.
7. Browser extensions based on Whisper
Several Chrome and Firefox extensions leverage Whisper locally or via API to transcribe and translate audio directly in the browser tab. They work on YouTube, Google Meet, and other streaming platforms without leaving the page.
Local audio processing eliminates the transfer of data to a third-party server, a point to consider since the tightening of GDPR rules and the entry into force of the European AI Act. Free services hosted outside the EU often remain vague about the retention period of audio files and their potential reuse for training models.
- Check if the extension processes audio locally or sends it to a remote API
- Control the permissions requested (microphone access, browsing history)
- Prefer open-source extensions whose code is auditable on GitHub
The choice between these seven tools depends on the source format (uploaded file or real-time stream), the target language pair, and the required level of confidentiality. For occasional use on short excerpts, ElevenLabs or Google Translate are sufficient.
For a recurring pipeline with technical control, Whisper via a no-code tool offers the best flexibility. Browser extensions are suitable for informal translations in video conferencing or video watching contexts, provided you check their data processing policy.