What you get
| Output | What it is | Use it for |
|---|---|---|
| Plain text | The whole transcript | Notes, articles, show notes |
| Timestamped text | One line per segment with its start time | Finding the moment something was said |
| SRT | Numbered subtitle cues | YouTube, Premiere, DaVinci, most video editors |
| VTT | Web caption cues | HTML5 video players and websites |
The transcript comes from Whisper large v3, a speech-recognition model that handles accents and background sound better than older engines. It is still a machine: it can mis-hear names, numbers and unclear words.
Limits that matter
- Up to 20 minutes per file. Longer recordings: split them with the audio cutter and transcribe each part.
- 1 free transcription a day per visitor; after that 5 credits each.
- No speaker labels: you get one stream of text, not who said what.
- Music with singing is hard for speech models. For a song, isolate the vocals first.
We do not store your audio. Your browser shrinks it to a small file, sends it over HTTPS to our API, and we pass it to our AI provider (fal.ai) for the job; the result comes back to your browser. Do not use it for recordings you are not allowed to share with a processor.
Tips for a cleaner transcript
- Use the original recording rather than a re-compressed copy when you can.
- Heavy echo or hiss lowers accuracy; run the recording through the voice clean-up first.
- Set the language when you know it: it avoids a wrong guess on short clips.
- Read the result once against the audio. Fix names, brands and numbers by hand.
Frequently asked questions
Is the MP3 to text converter free?
Your first transcription each day is free, with no sign-up. After that each one is 5 credits.
How long can the MP3 be?
Up to 20 minutes. For longer files, cut them into parts with the audio cutter and transcribe each part.
Which languages work?
Whisper large v3 recognises about 100 languages. The menu lists the common ones; Detect automatically works for most clear recordings.
Can it translate to English?
Yes. Choose Translate to English and you get English text from speech in another language.
Is my audio stored?
We do not store it. Your browser shrinks the file and sends it for processing; the transcript comes back to your browser. Our AI provider, fal.ai, processes it for the job.
How accurate is it?
Good on clear speech, weaker with noise, music, strong overlap or rare names. We do not publish an accuracy figure; always check important words against the audio.
Can it add speaker names?
No. The transcript has timestamps but no speaker labels.