Speaker 1: Hey there people, this is Chase Meyer, just wanted to show you a cool little automation I made to solve one of my own problems. One of the problems I keep having is I really want to have a way to record an audio file and then have that audio file, have something happen with it, other than just recording it and having it sit there and not being able to use it for other things. So this solution is really basic, I made it inside make.com. The first trigger here in the solution is to ultimately catch an upload, an audio upload. So when I record an audio file in voice memos on my iPhone and then I choose to share that to a specific Dropbox folder, I'm ultimately sharing it to this folder. And once a new file goes in there, this action, this trigger is watching the files to see what's new. Every 15 minutes this thing runs, actually I need to adjust that because it's burning a whole bunch of time. I really only want it to run from 9 a.m. to 5 p.m. maybe. And then we're going to have it go every day of the week and every month of the year. That way it's not doing it in the night time when I'm not really looking to utilize this stuff. So anyways, it's going to run every 15 minutes and if it finds a new file, it's going to bring it here, then that result will set some variables and I set variables so that I have a now year, a now date, and then a now time, and then path uploads, path process. So I have different things I can utilize further down the chain. After I set those basic variables, I download the actual data file. So that specific file uploaded, it's just a reference. And then when I actually download it, it looks like this. And that download file gets taken and sent to ChatGPT, we're using ChatGPT's whisper function which is transcription. So it takes that audio file, throws it to ChatGPT, and ChatGPT replies back with the entire transcription. How cool is that? And then it takes the transcription, it throws it to another ChatGPT module which is create a completion, where I train ChatGPT to create an output that is 8 words or less. And that output 8 words or less describes what the entire transcription is about. And then this, and I'll use this one for the file name ultimately. And then I have it transcribe again, I'm sorry, it takes the transcription and it formats it again, but it does it in a more advanced blog post style way. So it has very clear instructions not to edit my text, not to change it, but to go ahead and put header tags in the appropriate spots to kind of chunk it out. So it's kind of pre-formatted in HTML. And the advantage of that is that I can take this output and put it inside a document when I create it, and it's nice and structured. So as an example, we have, it goes through here and it searches for the yearly Dropbox folder to see if the 2023 folder exists, if it doesn't, it's automatically going to create the year. And then if the folder's found, it's going to go ahead and create the document. And it creates the document just like I would want it to be created with the custom title that has the date, the time, this one says transcription file, and then the eight character things with spaces, or I'm sorry, with dashes, hyphens instead of spaces. Then after the document's created, whether the folder was found or if it needed to create a folder and then create the document, it's the same thing. After it's created, it moves the audio file to the same folder and renames it. So that way, audio and the transcription are always together. Finally, if I, in the title of this audio file, if I typed post-worthy, then it's going to allow it to pass through this filter. And this filter's going to take the formatted content and it's going to rewrite it for like a post, like a LinkedIn post, to make it a little more user-friendly, and then it's going to post it on my LinkedIn account. So I can go from recording an audio file to ultimately having a LinkedIn post in like one minute. And that's a very cool thing. Still, clearly, there's going to be bugs every once in a while, but it's working very well so far. So here's an example. If I jump to LinkedIn and I go to my account, I'll look at my posts, activity, show all posts, and then here is the post that was posted. So I had one that I did yesterday that was an idea about building this automation. And I recorded the audio and I used that audio during testing, and then this is actually what it created based on all the audio. This is me thinking through the process of how the automation would work. The actual file that was saved is right here, and this was one I did from yesterday. So if I double-click this, this is the one that says, planning the transcription. And this is everything that was output from ChatGPT, which is really crazy. And then I have the actual audio file right here named the same thing, except it says audio file. So here's an example of that.

Speaker 2: I'm thinking about building this automation that would ultimately help me archive some of my thoughts in audible format, but then also in transcription format. So I think the automation I'm going to make is inside of make.com.

Speaker 1: So there you go. It's really cool. I thought I'd share it with you just in case anyone has interest in doing this themselves. I can ultimately share a copy of the automation I built, or at minimum, you know what, just share it with someone you think might appreciate it. All right, that's it. That's all I got for you. I appreciate you taking a look at this cool thing, and I'm going to keep making more fun individual mini-programs to solve my own problems that I will be sharing more frequently. Thanks so much.