Tools 4 All
Home
Additional Tools

Audio Transcription

Turn speech into text in about 100 languages. Drop an audio or video file or record your voice, then copy the transcript or download it as subtitles. The AI model runs in your browser, so your audio never leaves your device.

FreeNo sign-upRuns locally — files never leave your device

How to use this tool

Audio Transcription converts spoken words into text with small speech recognition models that run entirely in your browser.

How to use

  1. Add audio – drop an audio or video file (MP3, WAV, M4A, OGG, FLAC, WebM, MP4…) or click Record to use your microphone.
  2. Choose the language – pick the spoken language for the best accuracy, or keep "Detect automatically". The tool uses the model made for that language.
  3. Transcribe – click Transcribe. The text appears as each part of the audio is done; you can stop at any time.
  4. Save – copy the text or download it as TXT, SRT, VTT or JSON. Click a timestamp to play the audio from there.

Choose "Translate to English" to get an English translation instead of a transcript. It works with every language except Hebrew and English.

Audio

or

How to transcribe audio

  1. Add your audio

    Drop an audio or video file, or record from your microphone.

  2. Choose the language

    Pick the spoken language, or keep automatic detection. The tool picks the right model for it.

  3. Transcribe

    The text appears segment by segment, right in your browser.

  4. Copy or download

    Copy the text or download it as TXT, SRT or VTT subtitles, or JSON.

Private speech to text in your browser

Most online transcription services upload your recordings to their servers and charge by the minute. This tool works differently: it downloads a small open speech recognition model into your browser and runs it on your own device. Your interviews, lectures, meetings and voice notes stay private.

What you get

  • Transcription in about 100 languages, with automatic language detection.
  • Translation of speech in any supported language into English text.
  • Timestamps for every segment, and subtitles in SRT and VTT formats.
  • Audio and video files, or a recording from your microphone.
  • Small models, under 100 MB each, that run on any computer.
  • No upload, no account, no limits, completely free.

The models

Hebrew uses ivrit.ai's Zipformer (CC BY 4.0), English uses Moonshine Base (MIT) and every other language uses OpenAI's Whisper Base (MIT).

Key points are picked by multilingual-e5-small (MIT), only when you ask for them.

How to transcribe audio to text accurately

Audio transcription turns the speech in a recording into written text. It makes interviews searchable, lectures easier to study, podcasts accessible and videos ready for subtitles.

Tips for the best results

  • Choose the language: automatic detection works well, but naming the language avoids mistakes, especially in short or noisy recordings.
  • Use clear audio: less background noise and one speaker at a time give a cleaner transcript.
  • Use a computer: transcription runs on your device's processor, and a laptop or desktop is much faster than a phone.

Transcripts and subtitles

Download the result as plain text to edit or share it, or as SRT or VTT to add subtitles to a video on YouTube, in a video editor or in a media player. The JSON file holds every segment with its start and end time, for developers.

Who uses audio transcription?

  • Students: turn recorded lectures into notes.
  • Journalists and researchers: transcribe interviews without sending them to a third party.
  • Content creators: make subtitles and show notes for videos and podcasts.
  • Professionals: keep a written record of meetings and calls.

Need to record first? Try our Voice Recorder or Screen Recorder, or explore more tools.

Frequently asked questions

Which languages can I transcribe?
Around 100 languages, including English, Hebrew, Arabic, Spanish, French, German, Russian, Chinese, Japanese, Hindi and Portuguese. Pick the language for the best result, or let the tool detect it.
Is my audio uploaded anywhere?
No. The speech recognition model runs inside your browser, on your own device. Your audio and your transcript never leave your computer.
Is the transcription free?
Yes, completely free, with no sign-up, no minutes limit and no watermark.
Which model does it use?
One small model for each language, picked for you: ivrit.ai's Zipformer for Hebrew, Moonshine for English and OpenAI's Whisper Base for every other language. Each one is under 100 MB and runs on any computer.
Why does the first transcription take longer?
The first time you use a model, your browser downloads it once and keeps it in its cache. The next transcriptions with the same model start right away, even offline.
Can it summarize the transcript?
Yes. Click Summarize to get the key points: the sentences that best cover what was said, each with its time in the recording. A small AI model, multilingual-e5-small, picks them in your browser; it downloads about 120 MB the first time.
Can I create subtitles?
Yes. Download the transcript as SRT or VTT to get subtitles with timestamps for YouTube, video editors and media players.
Can it transcribe video files?
Yes. Drop an MP4, WebM or MOV file and the tool transcribes its sound track.
All additional tools →
What's My Headers See exactly what your browser sends to a server. What's My IP Find your public IP address, location and connection speed. String Extractor Extract readable strings from any binary file, like the Linux strings command. Unlock PDF Recover the password of a protected PDF in your browser.