Whisper to make SRT compatible text out

[enzyme69]

I tested Whisper commanline on Terminal, and really like that it's already have the timecode output with transcribed text and translation from the audio. However I wonder if it could be more of less SRT friendly:

Instead of: (CURRENT)

[00:00.000 --> 00:07.120]  okay this is a quick test recording of my voice in English just like I'm just
[00:07.120 --> 00:14.920]  gonna be whispering like random stuff and all the voice and the words that I say
[00:14.920 --> 00:20.200]  will be translated into another language and I will have a timecode I will be
[00:20.200 --> 00:29.400]  using this AI tools on the Mac OS it's called whisper and all I need to do is
[00:29.400 --> 00:35.400]  just to run a command in the terminal and it's gonna convert all the text okay
[00:35.400 --> 00:59.640]  cool

Maybe:

4
00:00:20,200 --> 00:00:29,400
using this AI tools on the Mac OS it's called whisper and all I need to do is

5
00:00:29,400 --> 00:00:35,400
just to run a command in the terminal and it's gonna convert all the text okay

6
00:00:35,400 --> 00:00:59,640
cool

[jongwook]

The SRT file is saved once the transcription finishes, while the progress is shown in the current format which has a more efficient use of the screen real estate. Is there a specific reason to prefer seeing the progress in the SRT format?

[petersterne]

Whisper produces TXT, VTT, and SRT files? I thought it only produced TXT and VTT files.

[enzyme69]

Ok you are right, I did not realize SRT and VTT and TXT output also generated after running whisper. Thanks! We can close this discussion.

[SaadBazaz]

Can we directly get SRT in the Python library instead of raw text?