Stream meeting to extract transcripts in real time

API and Webhooks
#1

Description
I am looking at ways to stream the meeting so that I can create transcripts in real time.
I wanted to check if it is possible to stream only the audio as I am not interested in video/screen shares.
I want to extract the test only.

Is there an out of box feature that can help with this ? Or do I need to stream the whole thing.

#2

Hey @Happy,

Currently the only way to do this is with the Closed Captioning API:

Thanks,
Tommy

#3

Hi @tommy, thank you for your response. My use case is different, I want to get the text of the meeting (in realtime) to do some analysis on it (for e.g. sentiment analysis) and present it to participants/host after the meeting.
I would like to get the information in real time so as to finish the task close to end of meeting.
Something similar to Otter.ai but a different use case. Hope that clarifies.

So the question is do I need to get the whole meeting streamed ? can i get only audio?
if its is streamed do you have sample of the type of data for e.g. how do I know if participant 1 is talking or participant 2 etc.

#4

Hey @Happy,

Right now you can only do this with the cloud recording audio, or live streaming.

You won’t be able to know which participant is talking based on the data in the livestream, but will with the cloud recording files.

Thanks,
Tommy

#5

Hi Tommy,

The cloud recording of the meeting is only available after the meeting ends right? The livestream metadata does not contain speaker details as you have mentioned above.

So can you suggest a way using which we can know which participant is speaking in real-time?

I am creating a service that uses the livestream data to do live transcription and uses voice recognition to authenticate the speaker in the meeting. But for that I need to know some authentic details of the speakers in real-time.

Thank you.

#6

Hey @Malay,

Currently this is not possible with the Zoom API.

Thanks,
Tommy

#7

Hello Tommy,

We are trying to extract speaker information from the cloud recording m4a audio files of a particular meeting. We couldn’t find any speaker info.

In the April 15 response above, you have written that we can find out which participant is taking using the cloud recording files.

Could you explain this part in a bit of a detail as to how can be use the cloud recording files to know which speaker is taking…?

It would be of a lot of help since we are stuck with this for quite long time.

Thank you

#8

Hey @Malay,

Happy to help! You can use the TIMELINE file to know which participant is speaking:

Thanks,
Tommy

#9