Audio to Text

Turn any recording into text

TextFromAudio takes an audio or video file and gives you the words as text. Drop in an MP3, M4A, WAV or MP4, press one button and read the transcript when it's done.

It works as an audio file to text converter for MP3, M4A, WAV, OGG, FLAC and AAC, and for recordings such as lectures, calls and voice memos. It also converts video files (MP4, MOV and WebM) to text.

You may also hear this called an audio transcriber or speech-to-text. Same thing. What you get is an audio transcript you can edit, copy or export.

The free mode has no limits: no cap on file length, on how many files you transcribe, or on languages. It uses lighter speech models, so it's a little slower and less accurate than Pro, and long recordings take a bit longer.

Transcribe an audio recording to text for free

To transcribe an audio recording to text, add your file, wait for the progress bar to finish and read the transcript on the page. Fix any wrong words, then copy the text or export it as TXT, PDF, Word, SRT or VTT. There is nothing to install and no account to make.

What you can do with it

• Edit the text right on the page. Fix a name or a number before you use it.
• Copy everything in one click.
• Download the text as TXT, PDF or Word, or as SRT or VTT for subtitles and captions.
• Use video files too. We transcribe the audio track, so an MP4 works the same as an MP3. Audio works in MP3, M4A, WAV, OGG, FLAC and AAC, and video in MP4, MOV and WebM.

Transcribe a video

A video is audio with a picture, so a video to transcript job works the same way: drop in the file and we transcribe its sound track. People use it for lectures, meetings, interviews and their own clips. MP4 files have their own page, and other video types, such as MOV and WebM, are covered there.

Supported languages

TextFromAudio understands about 99 languages, including English, Spanish, French, German, Portuguese, Italian, Russian, Chinese, Japanese, Korean, Vietnamese, Arabic and Hindi. The language is detected automatically, and the free mode isn't limited to any of them.

Accuracy is best for widely spoken languages and lower for rare ones, especially with the smaller models. A larger model gives better results.

Free and Pro

Free. No limits on length, files or languages. The speech model runs in your browser, on your own device. Your file isn't uploaded anywhere. The model downloads the first time you use it and your browser keeps it, so every later transcription starts faster.

Pro. Pro runs a stronger model on our servers, so it's faster and more accurate. Your file is deleted automatically as soon as it has been processed. The free mode stays as it is. See the Pro plans: 15 hours for $6 or 100 hours for $29, transcription only.

Who uses audio to text

Students and note takers. Following a lecture and writing notes at the same time is hard. Record it, turn it into text, then search it, highlight it and study from it.

Podcasters and video creators. A transcript is a quick way to get show notes, captions, blog posts or social media quotes out of an episode you already made. The SRT export gives you subtitles for the same video.

Interviews and meetings. Journalists, researchers and teams record the conversation, transcribe it, and pull out quotes and decisions without replaying the whole thing.

Voice notes. Talking is faster than typing. Record your ideas on your phone and turn them into text later.

Tips for a cleaner transcript

• Record close to the speaker. Distance and room echo hurt the result more than anything else.
• Cut background noise. Turn off fans and music, and skip the speakerphone if you can.
• Try to avoid people talking over each other. Overlapping voices are hard for any transcriber.
• Use the original file when you have it. A clean WAV beats a heavily compressed copy.
• Check names, brands and technical words. They are the most common mistakes.