How to retrieve transcription duration in minutes using Whisper with NodeJS and the OpenAI API?

Community
, ,
1

Hello OpenAI community!

I’m working on a project using NodeJS to interact with the OpenAI API for audio transcriptions using the Whisper model. So far, everything has been going quite well, but I have a specific question.

Is there a specific way to obtain the duration in minutes of a transcription performed with Whisper? I’m looking to integrate this information into my application, and I haven’t found a clear solution in the documentation.

I appreciate any guidance or suggestions you can provide. Thanks in advance!

1 Like
2

Welcome @gfbane23

When you say duration of transcription, do you mean timing of every segment within the transcription?

If so, you can set the response_format param in the request to "srt" or "vtt" in the boilerplate code provided in the API reference.

2 Likes
3

Hello @sps, thank you very much for responding to me, and yes, I am referring to the moment of each transcription, that is, the timestamp of the text it is generating. Something like second/minute (time) (transcribed text at that time). I will check the references you have given me, thank you very much! I couldn’t find anything in the documentation, I hadn’t checked the API reference, and I also read some examples where it was only achieved with external libraries like fluent-ffmpeg

Update, I tried the references you provided, and I was able to get what I needed. Thank you very much; sometimes, we get lost in the documentation, and the answers are right there. Thanks again. Next time, I’ll make sure to review the documentation properly.

1 Like
4

That’s what the dev forum is for. Feel free to share your project’s progress or ask questions.

PS: You can mark this as solved so that fellow community members can quickly find the solution.

1 Like