Speaker 1: You see these subtitles here? They seem to be accurate enough to make it look like I wrote it right? Well, what you see right now is not written by me, but it's written by an AI. Let me tell you more. So to give you a bit of a backstory, I've recently been trying to get back into making videos on YouTube, where I've started uploading a video about authentic reviews and if you haven't seen it yet, please check it out, I'd much appreciate it. And so whilst editing the video, I was thinking about whether or not I should have subtitles on the screen which you see in a lot of shorts and some long form videos. Ultimately, I didn't add it in the final video because I feel like you as a consumer should have an option to open those subtitles or not. So I've separated that so that people have the choice to turn it on and off, but I still had to generate it as an SRT file. And it was already taking me long enough to learn the interface with DaVinci Resolve as I was moving from iMovie to Final Cut Pro in a free trial of course. And then on DaVinci Resolve where I was thinking, there's gotta be a way to automate this. Now DaVinci does have their automated captioning, but it's paid, which being a small creator who's just trying to settle in the space, isn't ideal to make that jump yet. Then I did some research online, specifically on Reddit, as you've seen in my last video, and I came across a tool named Story Toolkit AI, which in theory does sound like a tool that would work. And after downloading and running that tool over my video, it works. Actually, it works really well. I didn't find a single mistake in my last video. I'm sure by the time I'm editing this, I won't have in this one as well. And best of all, the models are local. So let me show you how all of that works in practice. Alright, so this is the page to download Story Toolkit AI. I'll put it down in the description. And as you can see, it's a GitHub repository. It's not your typical website with a download button. It's a repository and you can go two ways about it. You can either download it directly, or you can clone it from the repository and then try to configure it yourself. It took me a bit to find a link to download the application, as I usually just grab the link of the repository and clone it through the terminal, which you can do, but I think you are better off downloading the stable release. So let's just try to find it here. So download setup installation and then we'll go to detail instructions. As you can see, it has instructions about installing the tool from source. So you can use the Git version or if you want to download directly, which is something I did, I just click this button and then you'll have this page talking about the latest release and you're not going to find the download button immediately. You're going to have to click this button and then scroll to the bottom where you'll have a zip file or an execution file, depending on which system you're using. Of course, I'm on Mac, so I'll click on the zip. I've already got a download, so I don't need to do that, but essentially you just run the installation and then you'll be good to go. All right, so now that story toolkit is installed. We've got a bunch of options. You'll be greeted with this new update screen depending on when you install it. So I'm going to press later. But you can see some options like ingest, open transcription, queue, search and assistant. I'm not going to dive too much deep into this, but since I want to focus on transcribing audio, I'll just go to file or transcribe audio files. So I just want to do audio files for this one. If you click on ingest, which is a typical route, it'll give you the option for video and audio, but it will process both. So you can't uncheck the video, but you can uncheck the audio. So it's a bit confusing. So I would just go and transcribe audio files. It just makes it much more easier. Don't have to worry about anything else. So let me open a file which I've edited before. All right, so now that I've got the file, you can see up here, it's from my last video, but it's the incomplete version, which will suffice for this example. So it's not too bad right now. So what you want to make sure is that transcribe audio is enabled and you want to make sure the task is transcribed. You could also translate it. I haven't tried it yet. So if you're running for a different language, then maybe you want to explore that option. Model medium. I've not had to click on anything else. So that doesn't really matter to me. Medium does fine for me. Device auto. And just in case I enable pre-detect speech and increase time precision, just to make sure everything works out well. Now you can change this if you want, but I didn't have to do anything. So it worked out fine for me. And then I'll click start and then this will load the model, transcribe it, and then we'll see how that all turns out. This will take a while, but when you first download it, it will download the model the application is using and then use it to transcribe the audio. So if it's your first time, you might want to take a break, go somewhere, or if your internet's fast enough, then it might be done pretty quickly. But not a bad time to get some tea or coffee. All right, that's done. So let's see how that has turned out. So we're going to close this and we're going to open open transcription. So it'll open a window to select a file. Click on the .json file that is produced. The .srt, you probably want to use it on your video editor or directly upload it onto your video platform. But let's click on the JSON. And as you can see, you've got the whole script. And if I put this in DaVinci Resolve, it will sync up perfectly. I am not joking. It syncs up really perfectly. So I didn't even need to do anything. I just chucked the .srt in and done. It works. So we can see how AI is making tasks much, much more easier in the video space. I'm quite used to it in the software engineering space, but it's pretty exciting to see this in the video editing world. And that's not going to stop. We're going to see more developments, more tools, more ease in terms of tasks that will help us in terms of achieving our end goal. All right, let me know what you think of this video. I know it's one of those one-off tutorials since I'm still learning video editing, but I look forward to improving my skills and bring any sort of value on this platform. So that's all and I'll see you later.