Build An Audio AI App Transcripts
Chapter: Feature 1: Transcripts
Lecture: Getting the MP3 URL

Login or purchase this course to watch this video and the rest of the course contents.
0:00 We've got all the supporting pieces in place. It's time to start using Assembly AI and actually transcribe this episode here.
0:10 So we're gonna need to do a couple things real quick. First, we're going to need to get a hold of the actual episode details
0:18 because the way it works is we're gonna have an MP3 file equals something like that. Then we're gonna pass that off to Assembly AI.
0:28 We're not gonna download the audio or provide the audio in any way directly to Assembly AI, although we could.
0:36 We're just gonna say, see that out there on the internet, go grab that and transcribe it. It can be things like MP3 files
0:43 or I think even YouTube videos. There's all sorts of stuff that we can put out there and say, go find that on the internet,
0:49 get it for yourself and transcribe it. So in order to make that happen, the first thing we need to do is say the podcast
0:56 is equal to, and we have this podcast service, podcast from ID and what are we passed in? A podcast ID.
1:07 So perfect, we'll say if not podcasts or podcasts is none, depending on how you wanna look at it, we'll say, right, it's not that we could do anymore.
1:16 This is done, so we're out. We also need the episode. And we need the episode by number. We'll have the podcast ID and the episode number
1:30 that we're gonna get. Now, these also are themselves asynchronous, so don't forget to await them because what you'll get back otherwise is a coroutine,
1:39 which is never none, but also doesn't work. It doesn't do anything. So there we go. So we have our episode. I'll have basically the same setup here.
1:49 Something like that. All right, so now we know we have our episode and the most important thing here is gonna be going to the episode
1:59 and saying called the enclosure, enclosure URL. And that's just RSS speak. And let's just do a little printout. Would transcribe. There we go.
2:17 So let's just go and hit that and try to make one of these jobs again. And we'll just say, let's consider doing episode 344 of Python bytes.
2:28 Gives us a job back. And what do we see? Would have transcribed. Oh, look at that. Python bytes, episode 344, AMA ask us anything from that URL.
2:39 Let's see if it works. Sure enough, there it is. Perfect. Look at that. So we're getting all the data that we need here.
2:47 We haven't yet sent this off to assembly AI, but we've got everything we need. The next thing to do is gonna be to use assembly AI
2:58 and import that and start calling functions like with the transcriber.


Talk Python's Mastodon Michael Kennedy's Mastodon