Gumloop’s Post

Transcript

Welcome to Jump. In this quick video, I'll cover the basics of flows and how you can automate your first workflow with AI using Gummo. Our goal is to take a website, understand what that company does in a one line description, and categorize it into the right industry. Write that to Google Sheets. Now what I'm gonna go ahead and do is add an input node. This will be the starting point. This is where we specify which URL we want to go ahead and grab O I'm going to call this. URL. And I'm going to put. IBM as the company, we're using Azure default value to test this for workflow. The next step is to go ahead and scrape that website, go to the website and grab everything on the home page. Now the magic of gum loop is you can pull information across nodes. So I can take take this value and pass it over as the URL we want to go ahead and grab. And if I run this, we're going to see the output what this flow does here on the right. So it's grabbing that URL. And then going to the website, now we can see the output. It is just a bunch of text, which is as a human really hard to understand, but as an AI it actually really is. And that's something they shine at. So let's go ahead and add another step here, which is asking the AI what it thinks A1 line description for this company could be. So I'm going to add a Ask AI step. Let me move some things out of the way like that. Let's write a prompt for what we want this step to do. So I'm going to provide A1 line description of this company based on its website. And I'm going to pass the website content from the previous step over as context for this AI to understand its job. So I'm going to run again. It's going to do three things. Grab the input, pass it over to the website scraper, grab the company's information and then get a one line description based on that information. We see here on the right. It's grabbed that one liner. The next step is to categorize that company as well. And again, this is something that. Guy does extremely well, so I'm going to add a categorizer step. And I can configure which categories the AI can pick from. So three options consulting. Technology. And finally. Manufacturing. Now note that I can even select which model is, give it a more powerful model here. And let's not forget we have to pass in the website content as input for the AI to understand what it needs to categorize. If I run that again, now we have 4 steps. Grab the input, pass it over as the website. We need to grab as the AI for a one line description and categorize across the three options. So here it's correctly picked technology and it's giving us A1 liner. Now note here at the bottom right. We see the credits. This is the cost of this flow. What's important to understand is that every node in gum Loop is free unless we have to go out to a third party like an open AI or anything else. A node only costs you something when it costs us something. So back to our flow. We've got our four steps. The final step here is to write this information to a Google Sheet. Again, we're going to go into our nodes, the building blocks of gum Loop. And let's look Google Sheet. We have some options. I want to write information. I'm going to bring that into my canvas. I'm going to pick a file. This is where I'm tracking the company information and gum loop. Automatically grabs the different inputs, the different columns in my sheet and just like before I can pull information from previous steps. Now the one liner is right here. The selected category is the industry. And in the input I'm going to run, grab the company URL and then this is going to go ahead and write it to my Google Sheet. Let's run the whole flow. Now we have 5 steps. Grab the input, scrape the website, get a one liner, categorize it and then write it to Google sheet. I see that it's run successfully and now if I look at my Google Sheet, I have that information there. So that is a first flow. In the next one, we'll cover things like looping. How do we do additional steps here and more complex workflows in gum? I'll see you there.