Lists
I broke it. That’s all
The AI Awakens
Successfully created the chat interface that actually calls on an AI (finally.) However, as you can see by the image, it has no history… not sure what to do about that. This might be a nightmare.
Also remember that login button I added… I’m starting to think that my go out the window… We shall see…
On a side note, Persona Forge (not v.2) is still down due to HackClub AI running out of credits. I tried moving to a free model but it hasn’t been working. It also is really slow. So I am waiting on votes on that… so once it starts working… then I will be able to get votes.
That’s all for now. I am about to start work figuring out the chat history. This is going to be funnnnnnn! (Wait… I might have figured it out in my head, but it involves a DB)
Wish me luck.
UI Updates
This devlog features the following:
- Cool Elastic GSAP animation on message added
- A Login button and send Message button that both do nothing
- The ability to send messages by pressing the enter key
- A cool blur effect for the input bar.
I am waiting on HackClub AI to get the credits back!
Started work on the chat area. The input does nothing, the messages are added via JS
Ship
going to ship it! Has some slight bugs, but I want to ship it and then sort out more bugs.
Pain down the road
I mentioned last devlog that moving my backend to vercel could cause some pain down the road. I was right. Now I have moved it back to being hosted on my end, so the preview doesn’t work any more. Sorry about that.
Attached is all the commits I made today.
Vercel Deployment
Now available in the public in beta!!!
https://character-creator-two.vercel.app/
Please give feedback via comments!
Good to go on Vercel. Realized though this might cause me pain down the road during development having the issue of committing to get the API to change. We will see though.
This is a smaller devlog for today but I want to go ahead and post it!
Added a system that limits users to 10 images a day, and automatically resets. IDK why I like it so much but I do!
Of course, there are many loopholes around it, but this whole project is mainly for my sister and she won’t know them whatsoever. I can count on that.
Now I just hope tomorrow it resets my request limit! 🤞🤞
Download Buttons & More AI
Well, so far this project makes 4 calls per image generated. That can’t get expensive, right? Right now it
- Checks if the prompt is a character
- Generates a list of traits from the prompt
- Generates a name for the file based on the prompt
- Generates an image
Recently I added a download button that downloads the generated image to your computer. The file name is auto-generated by AI which I think will be useful.
Next up is to figure out the share button.
I hope this project doesn’t blow up too much, because then I am going to burn through HCAI’s $3 limit so quickly.
I want to change the project name! Any suggestions?
What a productive 2 hours this was! Here is what has changed:
The things that work:
- Images are now being generated!
- The AI checker to check if the prompt is a character works, and will display a notification denying the request.
- You can generate a character all in the frontend, so technically this project is functional.
Things I need to fix:
- Images will take up the entire screen when displayed in the HTML Frontend. This isn’t supposed to happen
- UI is just garbage.
This project still has a long ways to go. (There is a roadmap on GitHub showing the completion progress.) Enjoy the picture of Superman holding up a car with the people in the car singing.
Have a foundation done so far. This includes the main frontend screen (just the opening screen), cool styling, and node.js AI stuff. Here is the funny part: It really hasn’t generated a single image yet in over three hours of work.
Did some outlining of the code, as well as got very frustrated at vercel!
Devlog features:
- Captcha start with loading symbol
- Some slight node.js so far for getting pictures from unsplash
