It's time for another blog post, folks.

It's August, that one month of the year which starts with depression and a lotta crying but ends with happiness and joy. Depression because August is when serious exams start, but then my birthday comes up and a couple more birthdays of people I know, and that's how it ends in celebration.

I've been lazy for the entire past month, except for the last few days of July. I was so lazy that I was even planning to publish a post called "I'm too lazy" but then I was too lazy to finish the post and ended up deleting the entry from my Notion table. That shows you the level of laziness I was experiencing.


For context, REHeader is an image header generator that generates a header image (kinda obvious) from the text that you input and other options. Jsoning, as some of you might know, is a node package that stores data in JSON format without the possibility of being corrupted.

I made REHeader within a few days of GitHub introducing really long Markdown descriptions for profiles, and shared it on a couple of places like Reddit, the Glitch Support forum and a few awesome lists on GitHub. I literally made it in a single day, using Vue within EJS and Express; and the code is so horrible (according to my current standards). The CSS used to haunt me, until a few weeks back when I fixed it to make it look more aesthetic.

Maybe as a result of sharing it on places more than Jsoning, and also as a result of it being added to awesome lists, it got more stars than Jsoning. And I'm kinda embarrassed about it. Not only is my most-starred repo a basically useless tool, it has really horrible code practices that would give most people nightmares.

I worked really hard on Jsoning, spent more than one day and yet some old project of mine gets the crown. To be fair, Jsoning isn't that bad, it got some pretty serious numbers on npm the other day.


On the bright side, I competed in a hack night by TinkerHub and won the best individual project! The task was to use a little something called Dialogflow API, a speech recognition API by Google that could do pretty interesting stuff with the input text, like detect the emotions in the input given to it. I used this emotion-detection stuff, technically called Sentiment Analysis, and displayed a meme based on the mood scores given to me. I had to code it from scratch within 5 hours. It was really stressful; I kept committing my .env file multiple times which resulted in me regenerating the API credentials more than thrice. The worst part about this was that I got no email about secret credentials being pushed to a public repo, contrary to how it's supposed to be - and so I only came to know about this midway through the project.

Happy friendship day to all my friends out there. That's right, I have friends!