Devlog 2

the project works now! i connected it to the hack ai (ai.hackclub.com). you can now connect multiply nodes between each other and chain them and everything and it generates something out of the input. uses google gemini flash models for speed but also quality.

The Idea (Devlog 1)

So I recently (like 2 weeks ago) I saw on YouTube a post about Freepik Spaces. I really liked this idea of having nodes that you can connect to something bigger and generating the result of 1 or more nodes with ai. like I head this vision in my head where I put in like an image and and text. that makes another image. with that I connect to I say more images. and then I have like multiply images and texts and idk and then I connect them to a presentation generation node. yea, I want to make like custom nodes. I think I can do it so the ai model gives back json that I can than put into the node to display it. So I think this is pretty cool just putting something in and it gets bigger and bigger and then there is a result, and then you can just change the beginning and the whole result changes!

What I did ( in almost 11 hours!)

I wanted to first make the node editor before doing any of the ai stuff. so I researched a bit on node editors react (because I use that for my project). I found react flow and its really like big and has many features. but mhh I was thinking its not that what I want. because like then I don’t really know how it works and how I can integrate my ai. so idk I just had the idea to make my own node editor. I started with a simple doted background. added moving, zooming. then I added my first node and made that you can move it to and it also scales and moves according. with the background canvas. then I added connections between which was pretty tricky to get right.

What I want to do next

Next Devlog I want to have integrated some ai like for text generation out of to text input nodes. I want to use hack club ai. which is I heard free for students.