PictoLexi

• 

  image generated

Inspiration

I'm useless in drawing for pictionary. I always end up confusing my team mate even more while trying to draw something on the board. So I decided to leverage recent applications of AI in image generation using prompts.

What it does

You just have to type out whatever you want to show depict as prompt, for example, entering Abstract colorful eyes, force Lines spray Hard watercolors Splatter brush human, haze, ultra-detailed, film photography, light leaks, Larry Bud Melman, trending on artstation, sharp focus, studio photo, intricate details, highly detailed, by greg rutkowski as a prompt might give something like this as an output

How we built it

I made a basic website for now using html, css and js. I also used Lexica.art's api to generate images using the custom query that the user entered.

Challenges we ran into

Getting the right image based on the custom query to work was tough. The api was returning random memes because it did not get the right format of requests initially. But later on I was able to debug it!!

Accomplishments that we're proud of

1. Managed to get the api to work
2. Managed to get the images to fit into the website without losing clarity

What we learned

1. Sending queries using apis
2. understood the parameters used in image generation like weights, seed value, guidance etc

What's next for PictoLexi

Right now it just works as an image generation tool based on prompts, but i need to add a way in which we can change the seed value etc. Also I have to try to make it into a complete game like scribbl.io with login links etc.

Built With

• api
• css
• github
• html
• javascript
• lexica-art
• replit