Speaker 1: Welcome back to Web Cafe AI, where we do daily chat DBT videos. In today's video, we're going to learn how to take an audio file, automatically transcribe it, use DBT to make sure that the wording is correct in the audio file, and then take it one step further and make a title and description based off that transcript. Now, I might look a little burnt, so leave a like if you feel like I need to wear sunscreen next time I go to the beach. But for now, let's go to jump into today's tutorial, where we're going to be looking at, essentially, a ZEP that I already created. I have already re-recorded this video, but sadly, my recording had paused halfway through, so I'm redoing the video. No big deal. So we're going to go ahead and start with a simple new file to folder, with the idea essentially being that when you create a video and you convert it down to an audio file, we're going to drag it in there and then begin the whole ZAPier. Now, if you're not familiar with DBT and ZAP, make sure to check out the video right there that goes how to implement chat DBT and ZAPier. But for now, we can go ahead and start this flow by getting our video. The video I'm going to be using is going to be a landing page video that I have, which is around 34 seconds long. If you're not familiar how to convert a video to an audio file, at least on Mac, all you have to do is come to export as. We're going to do audio only, and then I'm just going to rename this to GBT audio. And then from there, now we have the audio file we can use for this process. So I went ahead and uploaded that audio file to the folder that we're referencing in this app. Now, let's go ahead and start looking at how to use that audio file. So it went ahead and uploaded. I suggest using a audio file over a video file because it takes much longer to upload video files because they're so big. But for now, we're going to go ahead and do open AI. We say create transcript, continue, continue. And then for the file, you see, you can use video, but I really suggest using audio because it's just a lot faster. And then what we're going to use is the file variable from the file we just uploaded. And then the prompt is very special. So for me, at least an issue I've been running to when I'm transcribing my videos is when the audio picks up chat, GBT, it could misspells it a lot. So what I can do here is that any reference to chat, GBT should be spelled as chat, GBT. And then what you can do, obviously, is make a list of words that maybe you notice in your business, your personal videos, you know, gets misspelled a lot, which is annoying. But with GBT and Whisper API, we're able to really fine tune it so we don't have to go back and look through the whole transcript and reorganize it and basically re-write words that are misspelled. From here, you can choose your file format. I would suggest either text or SRT. I'm going to do SRT and then the language audio is going to be English. As you see here, they want to format it like this. And if you want to see how to format maybe in your specific language, just go ahead and click that link there. And you can kind of scroll through this Wikipedia page. But from here, all I have to do is hit continue and test action. All right, boom. As you see here, we have our transcript transcribed. And then essentially, you can see chat, GBT is spelled correctly. If there's any reference to Webcafe, sometimes that's not misspelled correctly. So I could add that to the prompt as well. But for the purposes of the video, we have our transcript. But now we're going to want to be able to push this transcript to something like a Google Doc so that that data is saved and we don't have to come back to this specific flow to find it, right? So we're going to do Google Doc. We're going to append a text document. You could create a new document from this text. And the purposes of this video, we're going to just use that event. We're going to choose the courses account. And then I have already created a document for this purpose, which I can find in my YouTube Zaps folder. And basically what I do essentially is you want to find the folder that is associated with the document you've created. If you're going through my route and then we can do is then find the document name. By doing this, you can't simply type in the document name because what's actually being grabbed by Zapier is in the document name. It's the underlying API reference token for them that they essentially are able to then grab the data that I'm about to push forward into the document. So then all you have to do is put in the text that was generated here and hit continue and then test that action. All right. It looks like it worked coming over here. As you see, it is not formatted perfectly. Obviously, we can go into the original document that we created over here and we can add HTML to this block in order to separate it and have it indented. I go over this a lot in other videos on my channel, so I encourage you to check out those videos. But for the purposes of video, we have gone ahead and grabbed the text or the transcript and pushed it forward to a Google Doc that can now be exported as XYZ. Now I want to show you something pretty cool. Let's go to take that transcript and take it one step further and create a YouTube title and description. So we use a chat. So we use a chat, a conversation, and then we're going to go ahead and put in the initial transcript. But first, we're going to do this based on this YouTube trend script. When you shift and enter, generate a YouTube title and description. And then, as always, we want to make sure we up our model here to GPT-4. So we're going to do that. OK, perfect. And then everything else we can keep as the default. We're going to continue here. Hit test action. And one thing I want to point out here is that I do have intentions here at WebCafe to launch a course that is very specific to YouTube, how to leverage completely automated YouTube channels, but also use YouTube as a way like stuff like what we're doing here. And basically input videos and get transcripts, automatic thumbnails, automatic descriptions, hashtags, keywords, everything you can think of. We do plan on doing that here at WebCafe, so make sure to subscribe. Stay tuned and get ready for a YouTube and chat GPT course to be released. But from here, as you see, we went ahead and got our transcripts, and then we were able to get our title and a detail based off that transcript. Now, I could go from here and format that title and transcript so we can separate it, remove title, remove description, then use it as a way to upload a video. But for the purposes of this tutorial, I just wanted to show you how to take a video, make it into an audio file, and then take that audio file, put it into Whispers API through OpenAI. And then on top of that, use chat GPT to leverage it to make a title and description. Feel like you've learned something. Make sure to check out that playlist over there. We go over a bunch of chat GPT and Zapier automations. If you like this kind of content on this channel, make sure to like, let me know, and put in the comments if there's anything else you want me to look at when it comes to these two apps. Maybe you want me to do more Gmail oriented Zapiers. Whatever it is, let me know. But without further ado, I'll see you in the next video.