Turn audio and video into text
Accurate AI transcripts with timestamps, in 90+ languages. Read along as the words appear.
- Fast
- Words in seconds
- Accurate
- Up to 96%
- Any language
- 90+ supported
- Files & bulk
- Uploads, playlists
Okay, so this is the chair. If you lean back, you can feel it rock.
Flip it over and press on each leg. The one that moves is the problem.
Getting the transcript
Opening the video
0:00Listening
That's the first hour.
Sign in with Google to unlockHow it works
Add a file or a link
Drop in a recording, such as a meeting, interview, lecture or podcast, or paste a link to a video, playlist or channel. Long files work too.
Read along as it appears
ScriptYT listens to the audio and writes it out with punctuation and paragraphs. The first lines show up within seconds while the rest is still being transcribed.
Copy or download it
Copy the text, or save it as a text file, a Word document, or SRT and VTT subtitles.
Why ScriptYT
- It listens to the audio itself
- Whisper Large v3 speech recognition writes out what is said, so it works on any recording, with or without subtitles.
- It reads like real writing
- Full sentences, punctuation and paragraphs, instead of a stream of short caption lines.
- It tells you where to check
- Lines the speech model was unsure about get a wavy underline, so you only need to review those.
- Nothing gets silently skipped
- Long stretches without speech, like music or pauses, are marked in the transcript so you know nothing is missing.
- The audio and the text stay in sync
- Click any word to jump to that moment, and the transcript follows along while it plays.
Questions
How do I transcribe an audio or video file?
Drop the file into the box at the top of this page, or select it to choose a file from your device. The text starts appearing within seconds, with a timestamp for each paragraph.
Can I transcribe a video from a link?
Yes. Paste the link into the bar under the file box and select Get transcript. Playlists and channels work too, with a paid plan.
Which languages are supported?
More than 90 languages, detected automatically. Accuracy is highest for widely spoken languages such as English, Spanish, Hindi, German, French and Portuguese.
How accurate is it?
On clear speech, Whisper's large model gets the vast majority of words right, including punctuation. Heavy background music, strong accents and crosstalk lower accuracy, and those lines are marked with a wavy underline so you can check them.
Can I download the transcript as subtitles?
Yes. Every transcript can be downloaded as SRT or VTT subtitles, as plain text, or as a Word document.
Do you keep my files?
No. Files and audio are deleted as soon as the transcript is finished, and they are never used to train AI models.
Which files can I upload?
Meetings, interviews, lectures, podcasts and voice notes all work. Use MP3, M4A, WAV, MP4, MOV, WebM and other common formats, up to 500 MB. Your file is deleted as soon as the transcript is done. If you are signed in, its transcript is saved privately in your history.
Can I transcribe a whole playlist or channel?
Yes, with any paid plan. Paste a playlist or channel link, pick the videos, and download every transcript in one ZIP.