Meme Rater

Finished the program you can view a demo here, shouldn't be randomized nor hardcoded, a bit iffy with the ocr recognition too but couldn't figure out the API for a another service so stuck with tesseract

Fixed all routes, brainstorming how to make this better, here is a meme you can use for the program

putting finishing touches on the app probably just a little while longer then ill be done, missing UI component, went wrong with routing somewhere missed something in the image

admitdatly used ai to center the top part meme judged and added that logo, makes for better visibility and appeal

forgot to define node version in json, making it so memeuplaoder tsx can be clicked on anywherein the box instead of that tiny space top right

making it better, responses still have randomness(too much at thatconst STORAGEKEY = 'memejudgemodelv2';
to keep everything stored in browser, ocr is kinda buggy with images mostly works with text

Testing memes relezing I hardcoded to much stuff. Trying to make stuff actually train saw when I did the same image randomness changed it too much

Made sure the app fully works on mobile and all brows3rs added aira labels and keyboard navigation focuses on ratios and text size and edited css cause title wasn't showing

fixed toast, notifications actually show up properly now

Fixed OCR functionality using useOCR hook wrapping Tesseract.js. This allows the app to extract and display any text found in memes, enhancing the feedback users get about their images. Tweaked performance and caching to reduce wait times for text extraction.

decided to go dark mode instead too lazy to add switcher

built out css, used some fomr other project set all routes in main tsx

building and mapping out css, very very very very very rough sketch

building ocr with tesseract, pritngin result in the console

Figuring out how to use vite beter and building the actual app, ocr coming soon

making config files

Hardcoded the app for now, making it actually analysize soon

Using tesseract for ocr cause I cannot afford google ocr, getting some ui components and design inspo

forgot to do devlogs, learned that there is tensorflow for js will be experimenting