How to use Audio to text
- 01Add a recordingMP3, WAV, M4A, OGG or a voice note, up to 30 minutes.
- 02Transcribed on our serverOnly the sound is sent. Nothing is kept after the text comes back.
- 03DownloadSave the text as TXT, or as SRT or VTT subtitles.

Speech in, text out.
Whisper, an open speech recognition model from OpenAI, listens to the recording and writes down what is said, with a time for every line. It runs on Cloudflare’s servers, which is much faster than a phone.
At a glance
- Price
- Free, with a daily limit shared by everyone. No sign-up.
- Privacy
- Only the sound is sent to our server for transcription. Nothing is kept.
- Accepts
- MP3, WAV, M4A, AAC, OGG, OGA, OPUS, FLAC, WMA, AMR, WEBM
- You get
- Text (TXT), or SRT or VTT subtitles
- Length
- Up to 30 minutes per file
- Works in
- Chrome, Edge, Firefox and Safari, on phones and computers
- Internet
- Needs a connection while it works
Questions
Is my file uploaded?
Yes, the sound is. Speech to text is one of the few Treeet tools that use a server, because speech recognition is far faster there than on a phone. Your browser takes out only the sound of the recording and sends it to our server, where Cloudflare Workers AI writes the text. We don’t keep the sound or the text. Converters and other tools work on your device.
Which languages does it understand?
Whisper knows 99 languages and finds the language by itself. For short or mixed recordings, pick the language in the settings. It can also translate the speech straight into English.
Is there a limit?
Up to 30 minutes per file. The tool is free, with a daily limit shared by everyone; when it’s used up, the tool pauses until midnight UTC.
How accurate is it?
Clear speech with little background noise comes out very well. Music, people talking over each other and poor microphones cause mistakes, so check names and numbers before you use the text.