shouryamishra011

I’m redesigning the UI again, making it much more user friendly. I’m also taking all the feedback I got and putting that in. I’ve added a language feature as well. I also have a sign in (using firebase) feature now.

I’m working on adding a “clarity pulse” feature, I’m also working on adding AI generated affirmations that take user data from entries and then proceed to take them and generate affirmations for the user.

I’m also working on getting my devlogs to be much much better.

But anyways, these updates are coming soon to ClearMind, you can check out ClearMind and MindAI at https://clearmnd.vercel.app.

Thank you for following along with me on this journey.
Welcome to ClearMindV3.

After locking in for so long, I present to you, HoloHand. Everything is functional, now time for you all to review it and test it!

So while I was trying to integrate the hand sensing and optimize it for slower computers, I realized that when on a faster computer, it wont use that computers GPU at all. SOOO I did a lot of research on how to work around this, I updated the model, I had to hard code some of the functions for the hand sensing using mediapipe and I had to learn all of mediapipes docs so that I’m able to continue with this. Now after doing all that I went back and started to code it, while I was coding it I decided to also add more blocks and shapes as there were only so many at the time. The updated version is almost ready for production, there is still a lot of work to be done but it should be done by tomorrow or the day after.

Also don’t mind the pic, it’s not supposed to look like that, it should resize depending on how big the users screen is but at the moment I haven’t fully implemented that feature. I have it only for the menu.

I locked the flip in and got a whole buncha stuff done.
So lets go over it:

• The UI/UX is fully done, might make a couple changes here and there
• The hand sensing is working (finally)
• Optimized it for slower computers as well
• Made it work so that you can use both hands
• Increased the awareness of the recognition system
• Increased menu sensitivity
• Added the grab, pinch features
  and there was a couple more things I did but those are less important. Overall HoloHand is almost done and ready for production soon.

I swapped devices (I’m working on it in school) and the menu is all messed up, so now I’m gonna have it resize based on the users screen.

I worked on the main UI features, I now almost have a functional app. Just need to get the menu functions working along side the actual tracking system.

Got the tutorial done! Now we’re gonna move onto the main UI.

I finished all the frontend changes! I believe its ready for production.

Spent a bit of time today working on the project, I got some more frontend redone. If you wanna see it, dm me @shouryamishra011 otherwise, wait till clearmindv2 drops.

Its so fried… i lit have a AP Physics test tmr.

Working on adding a whole bunch of AI based features as well as getting the mood logger functional!
Gonna change some of the front end next.

I’ve started working on ClearMindV2, taking all the feedback I received, I’ve completely redesigned ClearMind.

Meet, ClearMindV2.
Access will be dropping soon, I’ll be working on a couple more features before I release it though.

The app is done! I got all of the features to work, I think there might be a couple bugs but Idk… i need feedback for that… but yeah, the app is done!

Meet ClearMind.

I got a whole bunch of the AI features to work, I’m gonna call it a day here. I think I should be done with this project tommorow!

I got the all the pages and functions done! The overall structure is done, I finished most of the features, now time for the AI features.

Yo im ngl i lowkey reached flowstate in coding and got the AI sorting algorithm to work…
Basically what it does, it uses hackclub ai as the API running on gemini and it sends the info to the AI and the AI sorts it into like important, not important or medium and then allows the user to review it, then the user can talk about the task to what I call, MindAI, hence the name ClearMind. Theres a lot more features that I plan on adding like a breathing technique, personalization, etc.

Home page is done! Getting the features of it ready, along with a sign in page maybe.

I have added a shop feature. Users also now get 10 coins for every level they complete, in the shop user can buy skins. I also will add in the next update a powerup feature. I also made level generation much more smarter and added a better colliding algorithm to it.

I decided it would be best to give everyone access to creating their own levels. So I added a create your own level feature, where users can save those levels locally using browser cache and I also worked on the physics based movement; making it smoother. I have created 10 levels, users can create their own levels by either manually doing it or hitting auto generate. I also removed the coins and skins feature as I thought that I should focus on getting the game out there for now, and I’ll take feedback as I get it and if people want skins and coins and power up features, then I’ll add it. The image below is an image of how it looks, I’m basically done the game now. Time to submit!

Got 10 levels developed! The game is basically done now, I plan on adding a creator access so creators with the password can make their own levels.

Added the level generation and overall game concepts. Its basically done.
The level generation in Orbital Swing uses a procedural deterministic system.
Seeding: Every level ID (e.g., Level 3) acts as a “seed.” Because the randomness is tied to this ID, Level 3 will always have the exact same layout for every player, but it will be different from Level 4.
Iterative Placement: The generator walks along the X-axis (the length of the level). Every few hundred pixels, it places a Hook Point.
Path Logic: It calculates the height (Y-axis) of the next hook based on the previous one to ensure the level is actually “swingable” and doesn’t just put hooks in impossible locations.
Hazards & Loot: Between these hooks, the game “rolls dice” (using pseudo-random numbers) to decide if it should spawn a coin, a bouncer, or a hazard. As the level ID increases, the probability of hazards appearing increases, and the gap between hooks becomes more varied.

Got the home page down, working on level generation now.

Went ahead and fixed the cursor issues, I also added a part in the code where I can control how long it takes for the cursor to expand to the button size.

After multiple hours of learning and training. YA BOY GOT IT!

holy trial and error. this actually took so long.

Uh btw if ur a reviewer reading this… just try to keep the resume under 6000 tokens cus im broke as frick and dont have money to buy an api key… SOOO UHHH yea

HOL ON NOW WE MIGHT’VE JUST FOUND THE ISSSUEEEE!
The “Missing API Key” error is occurring because the ESM module system in the browser (used by index.tsx) does not automatically have access to a process global unless it is explicitly attached to the window object and referenced that way. While the shim was present, the build process on Vercel sometimes treats process.env as a static replacement target, which can lead to it being “undefined” at runtime if not handled correctly.

so uh yea… i got no clue how Im gonna get started on fixing this but hey I mean at least we know the issue now

I went thru and updated my api to a hackclub ai api key and yet it still isn’t working, I’m still trying to fix this issue and bud in the code. I’m new to this so learning this from scratch is pretttyyy tough.

But atleast the home page looks gas!

So working on the issue, I lowkey locked in for a bit n then it just stopped working… I have no clue how to get this to work. I’ve added the api key to my backend and my front end, pulling it from both ends but it still doesn’t work, I’m completley lost. If anyone knows how to do this then PLEASE reach out to me.

I tried to fix the AI issue, what i learned is that the API key is being hidden by vercel, and when I tried making it public it didn’t let me. I have no clue how to fix this issue, I’m completely stuck. I don’t think I can fix it, I googled a whole bunch of solutions but nothing was specific to this. If anyone has any idea, please hit me up!

in my other project folder, I merged these two projects together, so now when you open up R.AI, sign in, create your own AI then you can take a custom generated link and put it on your linkedIn or on whatever you want. For example: https://rai-builder.vercel.app/soumya-p-mishra-enterprise-leader

I also HAD to add a dark mode.

Got the AI end to work, I added a whole bunch of cool new features like a reactbits cursor and etc. I created a sign in page, a forgot password button, updated backend on supabase, I’m having trouble for vercel to allow me to let others use my api key so the AI doesn’t work on the deployed site which is an issue for review… and thats another reason I should make this a paid website, maybe like $5 per month to keep it running. Not too sure, im open to any suggestions though. Other than that the website is done! I’m fully done on front end and backend. Just need to add a paywall later on and thats it!

Finished the full home page, as I started to model the main AI framework, doing a lot of testing and etc on the AI framework right now. Below is an example of how it would look, I’m using my dads real resume. I also plan on adding a paywall because it costs money for the AI’s API

Got the live preview to work - its just a quick preview of what r.ai is and how it looks like

The Home Page is a professional storefront, designed with a “Bento-grid” aesthetic and glassmorphism. It uses a Live Preview Simulator to show recruiters exactly how they’ll interact with your digital twin before you even upload a file.
The Backend is a hybrid architecture: Supabase acts as your secure cloud vault (handling Auth and SQL storage), while Gemini 3 Pro serves as the “Cognitive Engine.” When you upload a resume, the backend doesn’t just store text; it performs a Neural Synthesis to map your career achievements into a high-dimensional persona.
This is the basis of Resume AI and it’ll only build from here.

Finished the website!
Added the portfolio page, and a email section.

I added a journey log with an animation on the line, this took me a while trying to get the thing to resize for phones without squishing itself, but in the end it worked out!

I added a “technical arsenal” showcasing my tech skills. Used React, Tailwind CSS, Framer Motion, GSAP. Took me a while to get the sizing right and the cursor to expand onto the following widgets but I got it

Got all the cards done! I also added some background particles, animations, and some other stuff like resizing to the users screen size.

Working on adding “cards” to the home page, when the user clicks on it, it’ll take them to my page.
I’m gonna do this for my linkedin and insta as well.

Using reactbits.dev i added a minimalistic animation to my name on the home page that when you hover over it, it expands. It took me a while to figure out how to translate the language from typescript to javascript but I got it, and then I also had to optimize it a lot because it was very laggy. It would constantly check where the users mouse is making it super laggy but once I optimized it to check per frame, it worked much better!

Created a home page and added a parralax animation to that, took me a while to learn how to do that! Now im working on making a keyboard that basically comes together and then the keys are like interactive so if I click on them then one of them will take me to my github, linked in, etc. Things like that. I want it to showcase my skills obviously. I am gonna animate this keyboard (pic 2) so that it builds itself on screen.

Got a sweet animation for the homepage done! They keyboard animation is giving me a LOT of trouble, I’m using react bits to learn some of the code needed for this

Working on the error from yesterday, I think something is wrong with my postcss.config.js file! I have no clue what the error is but I know I’ll fix it eventually…

I’m still working on getting the page set because right now i have just added some filler text and I see that the animations work but the page is broken.

Solving a whole bunch of error that I got from linking the UI pieces together. The front end of this will be running thru react, there really is no backend so I don’t have to worry much about that at all.

I got a whole bunch of errors that I went thru and fixed and now the page is running but the page looks broken so I’m fixing that now.

I’ve started to create my own personal website, im currently working on adding UI features as I want it to look good. Working on getting a nice user friendly interface rn!