Today I made a basic Landing Page also I discovered my Style Hope you will Love it.
0
Log in to leave a comment
Activity
Shipped this project!
“Cooked” is a fun web app that uses AI to roast your Spotify music taste. [Paste Public Playlist Link] I wanted to build something that wasn’t just a boring data tool, so I made this to be a savage critic that tells you why your playlist is “mid” or “basic.” One of my favorite features is the Language Sidebar. I realized that a roast feels way more real when it uses local slang, so you can now get roasted in Hindi, Spanish, Punjabi, and more!
How to use it:
Copy a link to any public Spotify playlist. [Demo Link: https://open.spotify.com/playlist/4C9lfeVFBeoWwAXfhpiJpu?si=0585798b877e479f&nd=1&dlsi=1757b732c2ae404d] Feel free to use yours make sure public playlist
Open the sidebar and pick your language.
Paste the link and hit “COOK ME.”
Check it out and see if your taste is actually as good as you think: 👉 Demo: https://cooked-six.vercel.app/
The Roast is Officially Live! 🚀
I finally hit a major milestone the backend is fully cooked and successfully talking to the frontend! Seeing the Spotify playlist data flow into the Groq AI and come back as a savage roast was such a proud moment for me. During testing, I realized that a standard English roast didn’t always hit the right spot, so I added a language sidebar. Now, getting roasted in your native slang makes the insults feel way more personal and hilarious.
Building this taught me so much about how APIs actually connect and why state management is so important in React. It wasn’t just about writing code; it was about understanding how different systems “talk” to each other to create a smooth experience. I’m feeling way more confident in my dev skills now!
0
Log in to leave a comment
I learnt about Spotify API through their Documentation which was very frustrating not gonna lie but I’m glad I was able to learn how to use it for this project. I just had to split the links text and use the text after playlist then send it via API to get songs of playlist. I also completed the FrontEnd with the help of Tailwind.
0
Log in to leave a comment
Shipped this project!
I’ve officially reached the finish line with Booktures! This project started as a simple idea for an AI-powered reader, but it quickly evolved into a complex challenge in engineering resilience and speed. My goal was to create a digital terminal that doesn’t just display text, but actively “sees” the story alongside the reader, translating paragraphs into high-fidelity visuals without the usual lag associated with AI generation.
The most transformative part of this build was mastering API interaction. I hit a wall early on with rate limits and slow response times from single providers. To solve this, I designed a “Parallel Racing Engine.” I learned how to handle multiple asynchronous streams simultaneously, firing requests to Groq, Gemini, and Hugging Face all at once. Building the logic to detect the “winner” and handle failovers taught me more about production-grade backend logic than any tutorial ever could. It’s no longer just a wrapper; it’s a resilient system that refuses to break.
On the design front, I took a massive leap forward by learning Tailwind CSS from the ground up. I moved away from messy, hard-to-maintain stylesheets and embraced utility-first design to craft a “Tech-Noir” aesthetic. I spent hours fine-tuning the split-screen viewport, custom scrollbars, and glass-morphism effects to ensure the UI felt like a futuristic piece of hardware. Seeing the progress from a broken layout to a sleek, responsive dashboard was incredibly satisfying. This project pushed my limits in both logic and design, and I’m hyped to finally share the working engine!
I’m excited to announce that Booktures is officially deployed and accessible to the public! You can check it out here: 👉 booktures-snowy.vercel.app
This update marks a major shift in the app’s architecture and stability. Here’s what went down in this sprint:
🌍 Deployment
The app is now hosted on Vercel. I chose Vercel for its seamless integration with the frontend stack, lightning-fast edge network, and incredibly easy deployment pipeline. Every push to main now automatically updates the live site.
🏗️ Architecture Shift: Client-Side PDF Parsing
Previously, PDF processing was handled on the backend. While functional, it introduced latency and increased server load.
The Change: I’ve moved the PDF parsing logic entirely to the client side.
The Benefit: By leveraging the user’s local hardware, we get near-instant results, reduced bandwidth usage, and a more private experience since the document doesn’t necessarily have to leave the browser for initial processing.
🛡️ Enhanced Reliability & Fallbacks
Moving to the client meant I had to account for different browser environments and potential processing failures.
Robust Fallbacks: I’ve implemented a multi-tiered fallback system. If the primary parsing method fails (due to a complex PDF structure or unsupported browser feature), the app gracefully switches to secondary extraction methods.
Error Handling: Users now get clear feedback if a file is corrupted or protected, rather than the app simply hanging.
2
Log in to leave a comment
Comments
Does that mean you are ready to ship?
Shiped!!
I’ve been working on this engine like crazy, and it’s come a long way from just being a basic PDF reader. In the beginning, I had it hardcoded to just draw one guy named Jax, but I realized that was way too limited. Now, I use Groq as a “Visual Director” to actually read the text and figure out what’s going on—like if it’s a person, a history lesson, or just a bunch of data.
The biggest headache was definitely when the Hugging Face servers started blocking me because I used the free limit too fast. To fix that, I built a “racing” system. Basically, I make the app ask both Hugging Face and Pollinations.ai for an image at the same exact time. I usually wait for the high-quality one from Hugging Face, but if that fails (which it does a lot lately), I just grab the Pollinations one so the screen never stays blank. I even added a 20-second timer so the app doesn’t just hang there forever if the internet is acting up.
I did run into a few bugs—like once I forgot to clear the timer and the app crashed, but I fixed that! Now, if the engine “overheats” or your Wi-Fi dies, you get a clear error message in a cool red box with a button to try again. It feels much more solid now, like a real tool instead of a buggy prototype.
0
Log in to leave a comment
In today’s sprint, I successfully overhauled the Booktures core interface, transitioning from a basic layout to a high-fidelity, “tech-noir” upload experience. I implemented a custom Tailwind v4 theme that leverages color-mix for dynamic surface depths and a neon primary-gradient for the brand’s “Ignite Engine” aesthetic. The centerpiece is a sophisticated, interactive upload zone where I designed a “half-in, half-out” floating button anchored via absolute positioning and a group-hover logic that triggers a synchronized purple bloom effect. By hiding the clunky browser-default file inputs and utilizing CSS variables for consistent border glows, I’ve created a tactile, premium UI that feels responsive and ready for the intensive PDF processing to follow.
0
Log in to leave a comment
Today was all about setting the mood for Booktures. We moved past the blank-slate phase and gave the app its soul, landing on a deep, cinematic dark mode that makes the colors pop like a neon sign in the rain. The upload screen isn’t just a utility anymore—it’s an invitation. We crafted a high-end interface where users can drop their manuscripts into a “digital fire,” trading boring buttons for a bold, interactive experience that feels more like starting an engine than filling out a form. The stage is set, the atmosphere is heavy, and the “kitchen” is officially prepped for the first chapter to start cooking.
0
Log in to leave a comment
I’m working on my first project! This is so exciting. I can’t wait to share more updates as I build.
0
Log in to leave a comment