Build An Audio AI App Transcripts
Chapter: Welcome to the Course
Lecture: Course Table of Contents

Login or purchase this course to watch this video and the rest of the course contents.
0:00 So what are we gonna cover in this course? We talked about the app that we're gonna build, but specifically, what are the chapters
0:07 and how are we gonna break up our time? We're gonna start out by welcoming you to the course. Congratulations, you're most of the way
0:13 through that actually. Then we're gonna be focusing a little bit on setup and making sure you can get this app up and running.
0:21 You know, make sure that you can get your editor configured to run the couple of commands that we're gonna need to, virtual environments,
0:28 getting the database running so that it just can chill in the background and do its magic. So we'll talk about that just to make sure
0:35 everything's good to go. Then I'll give you a tour of that starter code, that starter application. Remember I said we're not gonna start
0:44 from complete scratch because that would add a lot of time around something that's not really central to the idea of building audio AI apps,
0:54 although important, not central. So I'll take that starter code and we'll just spend a little time understanding
1:00 how all the pieces fit together so that you can jump in and kind of make the code your own as we go.
1:05 Then we're gonna add four core features to this application. Number one, the ability to generate transcripts using machine learning.
1:14 And we'll be able to do this from not just some files that we download, but anything on the internet that basically relates to a podcast
1:21 is gonna be really cool. So that's the first start. Feature number two is gonna be adding deep search.
1:28 Once you have all the information about the podcast, plus word by word content of what is inside
1:36 of that audio file, well, it becomes much more interesting to add search to that situation, right? So we're gonna add search.
1:44 We'll see there's some really cool things we can do to make that really fast and also really user-friendly using HTMX.
1:51 So this is gonna be a super fun part. Building on search, we were excited about what we got out of transcripts. What about summarization?
2:00 So we'll bring in some large language model features from Lemur at Assembly AI. And we'll be able to ask questions like,
2:08 give me a TLDR summary of this podcast, or give me action items, or, you know, it's open-ended. It's an LLM.
2:16 So we're gonna do a bunch of interesting things to get additional information on top of just the spoken word,
2:22 the show notes, and those kinds of things. And then we'll fold that, of course, back into feature two to enhance our search further.
2:29 Final thing, we're going to go and build a chat with the podcast episode. So imagine this, there's a podcast episode that interviews one of your heroes
2:39 or somebody you're a fan of, somebody who's really focused on some topic, you know, whatever it is out there in the world.
2:46 Wouldn't it be cool to kind of have a Q&A, open-ended conversation with them? Well, we're gonna come up with something along those lines
2:53 using LLMs and chats with all this information that we built up. It's gonna be pretty awesome. And that's it. We'll just wrap up the course
3:01 and make sure that you're good to go and build, you know, all set to go build whatever it is you want, armed with all this awesome knowledge.
3:08 So thank you for taking the course, and hopefully this roadmap looks exciting to you. It's gonna be a lot of fun, what we're gonna build.

Talk Python's Mastodon Michael Kennedy's Mastodon