GitHub Spark is a fun GitHub tool that, given a prompt, makes your front-end application a reality! It seems like a convenient, low (or near-zero?) code way to get something prototyped and published to the Internet quite rapidly. Check it out!
View Transcript
Oh my god, it actually worked. In this video, we're going to be checking out GitHub Spark to see if it's able to put together an application for us with just a simple prompt. Let's see what we can come up with. All right, so Spark is now available on GitHub, and I'm going to go ahead and start it. I have never used it ever before, so let's see what we can do. Going to press the try it now button. and we'll see where it takes us. My plan in this video is to see how GitHub Spark is going to compare with a Clawed Code setup that I did where I built a Pokédex. And in full transparency, I was not expecting Claw Code to be able to do it at all, but it ended up working perfectly. And you can check that video out up here
if you haven't seen it already. And that was using Clawude Code with Clawude Flow on top of it to have an agent swarm. But can GitHub Spark deliver the same thing? Let's find out. To be totally transparent, this is going to be a different implementation because we can see that it says create web apps with React and TypeScript. I was using a Blazer application, but we'll see if we can get something with React and TypeScript going. So, let's try it out. I'm going to be using the Whisper app, and this video is not sponsored by Whisper, but it will let me speak into my microphone, and that way I can get my prompt in here. And I'm using a foot pedal as well, so I can do this all handsree. I would like you to create a Pokedex application that allows me to track whether
or not I've seen a Pokémon or captured it. Use existing Pokémon APIs that are available on the internet. This isn't the exact same prompt that I used when I was talking to Claude Code, but they were both equally as sort of nebulous and arbitrary in terms of what I wanted. I specified in the other one for Claude that I needed a Blazer application. We're not getting that in this case. I was more specific here that I wanted to use existing Pokemon APIs. And for the Clawed Code one, I didn't even ask it to do that and it worked. Let's see what happens. Something that I'd like to call out is you can see that it's generating a PRD and then building the application. If you watch the clawed flow video, it actually demonstrates a very similar thing where it's creating a set of to-do items
and how it's going to carry out the design of the application. I think that this is a critical step when you're working with agents. I think that we'll see more and more of this as agent technology becomes more and more uh prevalent. You might have found that if you're working with an agent and not taking steps like this, it seems like it's kind of all over the place or it's just not delivering anything useful to you. And I think that this is one of the key critical steps. Oh my god, it actually worked. Okay, there it is. Super cool. Um, scene caught. Awesome. Let's scroll through. Very cool. Oh my goodness, the stats, everything. Just a reminder that if you're enjoying the video and find it interesting, give it a like and remember to subscribe to the channel so you can check out more programming
tutorials. How awesome is that? Okay, so I'm impressed. Okay, so the cloud code one, it worked as well. It made a Blazer application. This actually worked with like you you watched, right? There was I didn't type any code. I'm sorry. I'm still a little bit taken back by this suggestions uh listed on the left. So, this is cool. We don't get that from Claw, but I'm sure I could have just said like, "Hey, give me some suggestions and it could have come up with some ideas." Add favorite system to mark special Pokémon. Include Pokémon evolution chains and family trees. Add effectiveness. Let's try one of these. I don't know what we have to lose. Right. So, let me click this one. Um, add type effectiveness chart and battle calculator. Let's go. In the meantime, while that's going, I'm going to search for Mew. Right. So,
we get Mew and Mew. Oh, I guess this is refreshing as it's doing some things on the left, right? So, there's MW Mew2. Caught both of these bad boys. There we go. Very cool. This feels pretty snappy. It might not be easy for you to tell, but like when I'm kind of navigating. Very cool, right? The cards, like the slight animation, it's just it's really neat. We'll see what this can do next. And while that's going, I think that I can go over to these next tabs. We'll try it out. Like I said, I've literally never use this, so you got you got an honest reaction from me when the UI actually popped up and it worked, but we can pick a theme. Okay. Data, Pokemon tracking prompts. Okay. So, all of the prompts that get used, assets, I can upload things. Okay. So, this
is still going. So, that's cool. It didn't disrupt it as it was going through. Um, custom theme. It seems like I'm not able to select any of these while it's going though. Okay, now we'll update the main app. Let's go. Let's see if it did it. There's an issue with the battle calculator component. There better not be an issue. Okay. So, it says that it's now done. I can pick the battle calculator. Okay. Um, we're going to do let's pick something like Charizard. We'll do fire. Oh, interesting. This is duplicated. I don't know why, but you can see fires here twice. Maybe couple glitchy things to figure out. H. That's odd. Defending Pokemon Venusaur. Right. Very cool. So, we can figure out the effectiveness. Um, I don't know why we got the double fire part, right? Maybe that's just a bug that we'd have to
go uh I'm assuming we can ask it to fix that. We can get to the type chart. Right. So, this is just the mathematical calculation that it's going to be using over at the battle calculator. So overall, I think that this is super impressive. It's not perfect, right? And I think that this is the important thing to remember and I do want to constantly remind people this because I make a lot of video content, especially on my other YouTube channel, which is called Code Commute. And there are people that submit questions and I answer them in vlog format. And one of the things that comes up non-stop is people being fearful of AI. So yes, we just had AI, you know, using Spark in GitHub go and make a Pokedex for us. But the reality is like we saw if I picked Charizard attack type
like it's buggy. It's not perfect. It means that we still have to be able to interact with these things. A Pokedex, maybe there's enough examples of this kind of thing. We had an existing API that we could use. This might be something that is easier for it to interpret. and that's okay. But the reality is that when it comes to building more complex software, this kind of thing will be trickier and trickier for AI to get right. Of course, we're still early with AI. This kind of stuff will continue to advance, but I want to remind all of you that if you're a software developer, these are tools that we can leverage that enhance our abilities. It doesn't mean that we are replaced by them. With that said, I will continue to explore GitHub Spark. I am very impressed that we were able to get
something working from the beginning. I'm excited to see what else it has to do. So, thank you so much for watching and I'll see you in the next video.