How I Automated My Website With These AI Tools

Hey, what's up, Matt Wolf here. I thought it'd be really cool to make a video for you about what goes on behind the scenes over at Future Tools and how I actually add the new tools because the tools are actually using AI. So when you look at one of the tools and you read the description and you see all the content on one of the tool pages, I'm using AI behind the scenes to actually generate that. And I wanted to show you how I do that. So I have this spreadsheet here, my Future Tools dashboard. And I have a tab here that I call the Machine and it's called the Machine because it basically generates all of this content here just behind the scenes for me when I paste a URL. So here's what happens if I come here to this cell. I actually have a tool on my clipboard right now that I'm gonna go ahead and paste to this column here called moobert.com and I'm gonna wait a minute here and watch what happens. And boom, you see that appear there. And there you go. It just created this short link for me as well. Now, not only did it do that, if I jump over into my webflow dashboard where I actually manage all of these tools behind the scenes, I'm going to go ahead and refresh this here and boom, this Moobert tool was added as a draft. Now what I will do is I will click into this and clean it up a little bit. I'm going to go ahead and get rid of all this extra stuff behind the tool name. The URL slugs a little bit crazy. So I'm going to go ahead and get rid of all that. You can see it automatically pulled in this image and you can see it pulled in a short description. Moobert provides royalty free music and products such as Moobert Rendered Moobert Studio Mover API, blah blah, blah, blah, blah. And it's got some extra text. This was all generated from AI, so I usually come in and just kind of clean it up a little bit. I'm just going to retype this as AI generated music and then the long description. Mover is a platform for content creators, developers and brands to generate, create and access royalty free music. It offers products such as all of these and it offers a holiday deal of 23% off. I'm gonna go ahead and just delete that little section there. But that was all written by AI for me. And then down here I like to add a little bit of additional context to each one. So what I'll typically do is I'll double click my link here, I'll open in a new tab and then just sort of review the page so I can see how it works. Now for the pricing. I happen to know that this tool is a freemium product. You can use it for free. But in order to actually own the licensing on this product, you do have to use their paid plan here. So what I'm going to do is I'm going to go ahead and update the pricing model to freemium. I do have to do some manual updating. I do review every single product that I put through here and take a look at it myself and get some of these details. So for tags, I'm going to go ahead and put music down here under additional context and tags. I added a few additional things that people might search for that this might come up for. So let's do music, Generative music, AI Music Beats Songs, Sound bed. All right, and I'm going to go ahead and leave it at that. So now if anybody searches the site for any of these, this tool will show up for those search terms. And, and now I just go up here to the top right click Publish. Now if I go to FutureTools IO over here and sort it by newest to oldest, you'll see that Moobert is now in here. If I click in here, you've got a longer description. It's got the music tag, the pricing model, and if you click on here, it goes to my short link and redirects and automatically adds a UTM source. So if anybody's watching their Google Analytics, they will see that that it came from my website. That's the whole process of adding it. Now let me show you how that actually works behind the scenes. So that happens automatically because it is run on a timer right now using a tool called make over@make.com. now make is sort of like Zapier, but it's got more of like a visual flow to it. So if I come look at my scenarios, you can see I've got a few scenarios here that are running and I've actually turned them off and set them as run on demand so I can demonstrate how this actually works behind the scenes. So I'm going to paste two more URLs in here. Genius Sheets.com and Neutron Field.com these are two tools that were submitted through the submit form and I'm going to add them to the website. I have these scenarios here turned off so that I can walk through what they're actually doing behind the Scenes. What I'm going to do is I'm going to jump over to scenarios. This is my account. Here's how it works. So what it does is when these URLs are added to this spreadsheet, it goes into this integration here. And you can see how this is similar to Zapier, where it's connecting all of these APIs together. So what it's doing first here, let me go ahead and click into what it's going to do here. So it's watching this Google Sheet. So this Google Sheet that you just saw, it's watching for a new update to that Google sheet. You can see the sheet name is the machine. The sheet name here is the machine. It's just watching that spreadsheet to wait for something new to get added. And it can run up to 50 at a time right now is what I've got it set at. So, one, a new line is added to that spreadsheet. This little automation will start to run. And you can see down here that it typically runs every five minutes. Now, what does it do when it starts running? Well, the first thing it does is it goes through this API called Scraping Bee. So Scraping Bee is another premium product that I use. And what Scraping Bee does is it looks at a website and it scrapes all the text off the website. It just literally copies all of the text that's on the website and saves it into like a. A file in the background here and make. So you can see that it's grabbing from the input URL A. So if I look over here, this is A in my spreadsheet, it's grabbing from input URL A. I'm having it ignore images and css, because I don't need that for what I'm doing. And then there's this little bit of code right here that tells it to basically scrape the title of the website and scrape all the body text from the website. So that's what it's doing. It's scraping all of the body text and the title of the website from whatever URL I plug in right here using this Scraping Bee tool. So then once it scrapes all of the content from that website, this is where the AI really kicks in. So it uses the OpenAI GPT3 and it takes everything that was just scraped. You can see right here that it takes that body content. Remember I just showed you that it was scraping the body content of whatever website URL was there. It's scraping that body content and it's giving a prompt to GPT3 saying scrape this body content and then summarize what the above website does. That's how that description is provided. It scrapes the content, looks at everything on the website and then GPT3 summarizes what the site does. I have some extra functions in here. Max tokens 250, temperature 0.7. This sort of affects the length and the quality of what it generates from the GPT. So you can mess with these numbers. This kind of affects the length that it generates. This affects how closely it reacts to the prompt that you put in. So it takes the scrape content and then it summarizes what the website does and then everything it just did right here in that summary, it then goes to GPT3 again and it takes what was generated in this and it says describe the above in one short sentence. So now it's taking this long summary that it just made in this GPT3 run and it's generating one sentence out of it. Now these down here, these are error handlers. This is basically saying if there is an error in one of these, just go ahead and resume to the next file. You know, go ahead and skip that one and keep going because I don't want to list, you know, 20 items and then one doesn't work and then it just stops everything and nothing continues. So this is basically saying if one of the URLs on the sheet doesn't work during scraping B or one of it doesn't work during chat, GPT3 just go ahead and skip that one and move on to the next to just keep going. So we've got it scraped, then we've got a sort of longer description. Then we have it generating a one sentence description with GPT3. And then what it does is it re updates this original Google sheet here. So you can see down here sheet name the machine row number row 22. It's grabbing whatever row this was on. So you can see this here was row 80. So it'll grab whatever's in row 80 there. And then for the tool name it's grabbing the title. Remember we scraped the title back here from the scraping B. So under tool name it's plugging the title in here under slug. This is the short URL that it's going to use. Basically it's got a little script here that removes the HTTP, it removes the dots, it removes slashes and the wwws, and it removes all that stuff and then just creates sort of a cleaner looking slug for the URL. And then down here what it's doing is it's pulling in the URL of the image. So if the site has an image like this one, had this image here, it's automatically pulling in that image from the site and placing it into column K. And then down here it's taking our short description that we generated in this second chat GPT generation here, and it's plugging that into the short description. And then under long description, it's taking the original description that we created here and that's the end of this process. And then I have a little, you know, sleep for 30 seconds here because I don't want it to run this OpenAI GPT3 over and over and over again at high speed because if you do too many requests per minute, then it's going to shut you down and stop working. So I have a little timer here, so it'll run through this process, wait 30 seconds, go to the next URL, wait 30 seconds, go to THE NEXT URL, wait 30 seconds. So I just plugged these two URLs in here. Genius Sheets and Neutron Field. Now if I go ahead and click Run once, it's going to go and process both of those. So it looked at my Google Sheet, saw there was an update, scraped all the content on the website, now just converted it into a description, converted it into a shorter description, and now it updated my Google Sheet. Boom. You can see there's the title from the website, there is the slug that it created, there is the image URL that it scraped, and here is the long, the short description that it created and the long description that it created. Now you can see it's going through its little 30 second wait period before it scrapes the next one. And that one will start any minute now. So you can see. All right, looked at Google Sheets, saw another one, scraped it, ran it through GPT3, ran it through GPT3 again, updated the spreadsheet, Boom. You can see that once again, pulled in the title, pulled in the slug that we're going to use, pulled in the image, pulled in the short description, pulled in the long description. All of that was done through AI in the background. All right, now what about this short link? For the short links, I'm actually using a tool called Link shortening. It's at short IO and you can see here that it created that Moobert short link for me and it automatically added this UTM source here. So how did this short link get created and how did this tool automatically get added into webflow? Well, once this flow finishes, it actually goes into a Second flow here, which watches the same Google Sheet to see when there's been an update. So you can see this is the same Google Sheet, the machine, same limit 50. So it'll do it up to 50 at a time. And then what it does is it's got an API connection into this short IO platform here. So it looks for the input URL here, which was the line that I plugged in. In line A, it looks for the slug that we created, which is right here. So it's looking at that line, which is line C for the slug, the link title in the dashboard here, that's the title that it puts here. What it's doing is it's looking in column B for that, which is our title column that we generated earlier. And then it's adding a UTM source of futuretools IO automatically. So you can see when it redirects, it adds this UTM source. So if somebody's looking at their Google Analytics, they'll see where the traffic came from. Then what it does is once again, it updates the Google sheet. It updates the same row that we're on. And then under main tool URL, which is column I, it'll update this with the new short link that was just created in this short IO tool. And then once it's done updating that Google Sheet, what it does is it updates all the information inside of webflow. So inside of webflow, it plugs in the main tool URL, which is this short IO link that we just created. It plugs in the tool image that we just scraped. It plugs in the short description, the long description for sorting. It actually uses the slug again for the sorting. That's just a weird webflow thing where it needs to be all lowercase for sorting, because it's sorts the lowercase and uppercase stuff separately. So I basically made a separate sorting line so that it sorts by that. If there's any relevant videos, I like to plug them into column W over here. And then it'll automatically pull that in as a relevant video. And then for the tool name once again, it pulls in line B, which is the tool name right here. And that's everything that it's doing in the back end. So if I go ahead and click Run once here, you'll watch it kind of crank through these here. I don't have to put any sort of time delay at the end because this one can just move fast with no issues. And if I jump back over here, you'll see that it updated my URL with my Short link here I click on Future Tools link Genius Sheets. It will automatically redirect to genius sheets.com with the UTM source future tools already built into it. Same with this one. Nerofield.com created a short link. If I click on it, it'll redirect to the website with my UTM already built into it. Now if I come over to Webflow here, go ahead and refresh. You can see it added these two tools as a draft and I'm going to go through that same process of reviewing them, clicking through them, checking the price, seeing if I need to clean up any descriptions or tags, and then adding them to the site. So I legitimately look at every single product that goes on the site. I click around, I figure out the pricing, I figure out what sort of tags it should use so that the search is more contextual and people can find what they're looking for with the search. And I clean them all up. Now I'm adding about 20 tools per day every single day. And the whole process takes me about maybe 30 minutes every single day. But you can see that if I didn't use AI in the background and I didn't use these no code automation tools like make.com, then it would probably take me hours, if not all day to update all of these tools. So AI is awesome. And that is how everything is working behind the scenes. That is the Future Tools back end that's running on Autopilot. I literally just plug in URLs right here. My machine over here on make just goes and does all the rest for me. And then once it's done all that, I jump into webflow and I clean things up a little bit. So hopefully somebody out there found that helpful. I thought it was pretty cool, especially when I discovered that make.com tool, I really like it. It's sort of like zapier with this visual flow and it connects with so many APIs. Like hundreds if not thousands of AP. Like if I was to go and add a new module here and click on search applications and just type like anything. Like, look at this list of APIs it connects with. It's insane. So pretty cool stuff. I love this kind of stuff. Hopefully other people nerd out about this as much as I do. And thanks for watching. See you guys in the next video. Check out FutureTools IO.