DEV Community: Yash

I Built a 'Perfect' Niche App with Zero Backend. Learned More About Product Prioritization Than Any PM Book

Yash — Thu, 30 Oct 2025 17:24:38 +0000

I watched students spend six hours on topics worth 0.5% of their final score, and founders spend six weeks on features used by 0.5% of users. Same problem, different domain. So I built Tracker for GATE Rankers (ECE) — a weighted progress tracker that turns study time into intentional effort.

Core Idea

Bundle the syllabus in a single syllabus.json, persist progress in localStorage, and calculate a client-side weighted score. Completing a high-weight topic instantly feels more meaningful.

Why It Works

Zero signup friction & offline-first
Privacy-first (data stays local)
Instant performance (no backend)
Weighted progress = smarter motivation

Single Code Block

const saveProgress = (topicId, completed) => {
  const key = 'gateProgress';
  const progress = JSON.parse(localStorage.getItem(key) || '{}');
  progress[topicId] = { completed: !!completed, timestamp: Date.now() };
  localStorage.setItem(key, JSON.stringify(progress));
};

const calculateWeightedScore = (subjects, userProgress) => {
  let total = 0, achieved = 0;
  subjects.forEach(s => s.chapters.forEach(c => c.topics.forEach(t => {
    total += t.weightage;
    if (userProgress[t.id]?.completed) achieved += t.weightage;
  })));
  return total ? (achieved / total) * 100 : 0;
};

Key Lessons

Show weighted value for every task.
Visual progress > raw counts.
Remove signup friction.
Privacy builds trust.
Beautiful UX is a retention moat.

Tech Stack

React + TypeScript | Tailwind CSS | localStorage | Static JSON | Deployed on Vercel

Building this taught me more about prioritization than any PM book. Weighted effort beats equal effort.

webdev #javascript #react #product #buildinpublic #showdev

I got fed up with overpriced, limited AI wrapper, so I built a free, open-source one.

Yash — Thu, 21 Aug 2025 11:49:57 +0000

The Late-Night Prompting Problem

Picture this: it's 1 AM, you're deep in a coding session, and you're stuck. You have five browser tabs open:

One for GPT-5 to help refactor a tricky function.
One for Gemini 2.5 Pro because it's better at explaining complex concepts.
One for a specialized coding model like Deepseek.
And a couple more just in case.

You're constantly copy-pasting prompts and responses, trying to stitch together the best possible answer. It’s a mess.

Then you see ads for these slick, YC-backed "all-in-one" platforms. They promise to solve this exact problem. You get excited, you sign up... and then you see the catch. They want a hefty monthly subscription (₹999 in India, which isn't trivial), and worse, they slap you with a tiny token limit.

It feels like they're selling you a Ferrari but only giving you one litre of petrol. For a developer, that's just insulting.

My Solution: No BS, 0 Cost, No APIs

Fueled by some late-night frustration and a lot of chai, I decided to build my own solution. It’s not a "startup." It’s a tool.

I call it AI Beast-A

It’s a dead-simple web app with one job: you enter one prompt, and it hits the APIs of four different top-tier models simultaneously. You get the results in four clean, side-by-side columns.

No marketing gimmicks.
No "magic" buttons that are just prompt += "please make this better".
And absolutely no token limits.

It's built with a simple stack (SvelteKit, Tailwind CSS) and does exactly what it says on the tin.

Why I'm Keeping it Free & Open-Source

My philosophy is simple: as developers, our job is to break down walls, not build new, expensive ones. I'm funding the API costs myself for now because I believe a tool this fundamental should be accessible to every student, freelancer, and developer in India without a paywall.

It's an honest tool for a community I love.

I'd be genuinely stoked if you guys could check it out, try to break it, and give me some honest feedback. And if you like what you see, the whole thing is open-source.

Live App: AI Beasta

GitHub Repo: go ahead, star it if you feel like it 😉
Let me know what you think in the comments. Cheers!

I Discovered a ‘Secret’ Citation Tag in My AI Answers—Here’s Why It Matters

Yash — Mon, 04 Aug 2025 08:59:32 +0000

Yesterday I was polishing a Dev.to draft when I spotted something truly bizarre in my Markdown:

text
:contentReference[oaicite:0]{index=0}
No human would type that. It looked like a sneaky citation placeholder that never got stripped out. 👀

The “Filter” Nobody Told Me About
What I saw: A raw tag instead of a footnote link.

What’s actually happening:
Every AI-sourced fact is stamped with an internal marker so editors (or automated “moderators”) can verify sources later.

A Markdown–HTML filter is supposed to replace these with tidy [1] footnotes… but mine slipped through before that step ran.

How I Uncovered It
Step 1: Wrote up my review of Hateble.dev

Step 2: Hit “Preview” and spotted the odd :contentReference[…] tag in the live render.

Step 3: Traced it to the AI’s citation system—basically a “belt-and-suspenders” audit tool that never meant to show end readers this raw syntax.

Why You Should Care
Transparency in AI writing. Knowing how and where sources get tagged helps you trust (or challenge) any bold claim.

Hidden metadata risks. If left unfiltered, you might accidentally expose internal tooling details—or worse, break your own Markdown.

Conspiracy fuel? It’s tempting to say “they” are spying on your clipboard. But the real culprit is just a busted filter.

TL;DR
I found a secret AI-citation tag in my draft.

It exists for audit purposes and should have been auto-removed.

If you see it too, add a quick find-and-replace, or fix your rendering pipeline!

I Stumbled on Hateble.dev AI ‘Roast’ Tool for Your Website (But It’s Throwing 500s)

Yash — Mon, 04 Aug 2025 08:45:08 +0000

“Don’t make it ❤️ Hateble.”

— the very first thing you see on Hateble.dev

Last night I was ego-searching my own domain (don’t judge 😊) and a wild result popped up: Hateble.dev.

The promise? “The brutally honest AI website analyzer to make your product ready to market.”
Naturally, I clicked. Here’s what happened, who (I think) built it, and why I’m still cheering for it even though it’s currently throwing a nasty HTTP 500.

First impressions

A gorgeous orange-to-black gradient hero.
A big input box for your-probably-broken-site.com.
A bright “🔥 Analyze my site for free” button.
Under the button: “HTTP error! status: 500” — every single time. Ouch.

UI-wise it oozes the same heart-shaped logo and typography you’ll find on Lovable.dev, the well-funded “build apps by chatting with AI” platform. :contentReference[oaicite:1]{index=1}

That immediately set off a branding déjà-vu…

Who built Hateble.dev?

There’s no /about page, no footer credits, and WHOIS privacy is on.

But a few breadcrumbs point straight at Lovable’s orbit:

The heart logo. Hateble’s word-mark uses the identical ♥ glyph Lovable uses for its brand.
“Works with any vibecoding tool.” The bottom of the page lists bolt.new, Cursor, Lovable, and Replit — exactly the roster Lovable’s docs highlight when they talk about “vibe coding” tooling. :contentReference[oaicite:2]{index=2}
Same color palette & type scale.

My best guess: Hateble is a tongue-in-cheek side project by someone inside the Lovable team (or at least by a designer/dev who hangs out in their Discord). Think of it as the sarcastic sibling: Lovable helps you ship, Hateble tells you why what you shipped kind of sucks.

What it says it does

Runs an AI/LLM agent against your live site.
Spits out a blunt UI/UX + performance audit (copy, layout, CLS, Lighthouse, etc.).
Gives a fix-list “so your launch isn’t cringe”.

If it sounds like ChatGPT + Lighthouse + a roasted Dribbble critique mashed together, that’s probably the stack.

Why I still like the idea

Negative feedback ≠ bad vibes. Early-stage founders need harsher truth, faster.
Automated QA for vibe-coded apps. Lovable users spin up full apps in hours; an instant “roast” bot fits that workflow.
Fun branding. Hateble is the perfect villain pairing for Lovable.

Lessons for indie builders

Takeaway	Why it matters
Ship a landing page early	Even broken, the site captured attention (I’m literally blogging about it).
Brand extensions work	The “evil twin” trope is memorable and shareable.
Error messages are UX	A plain 500 in red text kills hype fast. Consider a witty fallback (“Our AI passed out from all the roasting — try again soon!”).

What’s next?

I dropped my email into their (working) wait-list form — fingers crossed.
If/when the API comes back, I’ll run a full teardown on my own site and post the results.
Meanwhile, you can DIY something similar with OpenAI’s JS SDK + Chrome-Lighthouse if you’re brave.

TL;DR

Hateble.dev is a (currently broken) AI tool that promises to roast your site so you can fix it before launch.
All signs point to it being a cheeky side-project from the Lovable.dev ecosystem.
Even with a 500-error wall, the concept is 🔥 — and a reminder that honesty (and a bit of snark) is sometimes the fastest way to a better product.

Have you tried Hateble yet? Did you manage to get past the 500? Let me know in the comments — or roast this post, I can take it.

Happy shipping, and may your next deploy be more Lovable than Hateble. 🧡

How to Sound Like You Have a Philosophy

Yash — Mon, 28 Jul 2025 08:34:13 +0000

Begin with an idea.
Begin again.

Make it sound important.

Then important-er.

Say something vague that sounds like a quote from a misattributed philosopher but ends abruptly with a

Pause.
Thought.
Clarity.

Then we get really serious

By adding capitalized Ideas.
By capitalizing more Ideas.
By Capitalizing Literally Everything.

And end it all on a confusing yet powerful note:

Cabbage.

Building an AI Test Simulator for JEE/NEET – Need Help With Pricing and Feedback

Yash — Mon, 28 Jul 2025 03:17:01 +0000

Hey Devs and Builders 👋

I’m working on something close to my heart — an education app for India’s toughest exams: JEE and NEET.

We’ve built an AI-powered tool that can:

*Generate 100% accurate test questions (no randomness)
*
Let users choose specific topics they want to practice

Simulate a full test environment (timer, OMR-style, question switching, etc.)

Provide deep analysis after the test, like time spent, weak areas, and detailed solutions

It's like having your own personal AI-powered coaching assistant.

We’re also building a second tool:

A Feynman-style question solver. Upload a physics/math question (even as an image), and it gives a full conceptual breakdown and solution — clear, simple, human-style explanations.

**What I Need Help With:
We’re in early testing and want to understand:

💸 How would you price this?

Monthly?

One-time payment?

Freemium + premium features?

📦 Would coaching centers or schools want a bulk version?

🧪 Would you personally pay for an AI tool that helps you understand, not just answer?
**
This is built for India first, but we think the concept has global potential.

If you’re a student, parent, dev, educator, or just someone passionate about education, I’d love your thoughts.

Let’s build something that empowers real learning, not just rote performance.

Thanks 🙌

SoloStack: The Toolkit for Solo Founders (Share Your Story with Us)

Yash — Fri, 25 Jul 2025 15:47:34 +0000

Built by Solo Founders, for Solo Founders

Hey, I’m Yash — a solo founder who’s been exactly where you are. I created SoloStack to pay it forward, saving you the time and money I wish I had when I started. Let’s build something amazing together. 🚀

As a solo founder, my journey from idea to a profitable business was overwhelming. From finding the right tools to managing every aspect of business, it’s easy to get lost in the sea of options.

That's why I created SoloStack, a curated directory of tools designed specifically for solo founders. It's the toolkit I wish I had when I started, and it's here to help you avoid the expensive mistakes and countless hours spent researching.

Visit SoloStack

What You'll Find 🌟

Here’s a sneak peek of what you can expect:

127 Curated Tools across 12 categories to help you grow your business from idea to scale.
45+ Free Tools to get you started without breaking the bank.
Success Stories of real solo founders who’ve used these tools to build $50M+ in revenue.

Features of SoloStack:

1. Getting Started

A roadmap for building your business without overspending on tools.

2. Industry Stacks

Tailored tool combinations for different business types (e.g., SaaS, marketplaces, etc.).

3. Tool Comparisons

Filter tools by pricing, category, and stage of your business journey.

Success Stories 📈

Here are some of their journeys:

Pieter Levels (Nomad List): Built a $500K+ business using GitHub, Stripe, and Mailchimp.
Justin Jackson (Transistor.fm): Co-founded a 7-figure SaaS company with the help of these tools.
Danny Postma (Headlime): Built and sold an AI copywriting tool for 6 figures.

Why It’s Different:

No Affiliate Spam: Just the tools that work.
Completely Free to Explore: 45+ tools at no cost to you.
Curated for Solo Founders: I know the struggle, and I’ve made sure this toolkit fits the needs of solo entrepreneurs

Join the SoloStack Community 🌍

Building something amazing as a solo founder isn’t easy, but with the right tools, it’s possible. SoloStack is here to empower you to build without limits.

Start your journey now: Visit SoloStack

SoloStack — Empowering solo founders to build without limits. 💡

A Little Scroll Through Time ⏳

Yash — Fri, 25 Jul 2025 15:22:13 +0000

Sometimes I feel tired.

Sometimes I want to feel grateful.

Sometimes I just want to be amazed.

So I made a small web app — where scrolling = time passing in a day.

And as the hours go by, you see how much has already happened.

Flowers bloomed
Hearts beat
Babies were born
The world moved forward

You can scroll through the day, from 00:00 to 24:00, and watch the numbers grow smoothly. It's not a productivity tool. It's not a dashboard. It's a little window into the massive beauty of everyday life.

👉 Try it here
I hope it brings you the same quiet awe it brought me while building it.

Let me know what you feel after scrolling 🙏

Built with love, numbers, and a tired but curious heart.

Automated Crisis & Opportunity Response System

Yash — Sun, 08 Jun 2025 12:23:13 +0000

This is a submission for the Runner H "AI Agent Prompting" Challenge

What I Built

I developed an intelligent automated system that continuously monitors global news feeds, social media trends, stock market fluctuations, and industry developments to instantly identify potential crises or opportunities. The system automatically drafts response strategies, contacts relevant stakeholders, creates contingency plans, monitors brand reputation in real-time, and executes predetermined actions based on severity levels—all without human intervention.

Click here or image above to open DEMO video

Core Problem Solved: Traditional crisis management is reactive and slow. By the time organizations recognize emerging threats or opportunities, valuable response time has been lost. This system transforms crisis management from reactive to proactive, enabling instant strategic responses.

Demo

The system successfully completed a comprehensive monitoring cycle, delivering:

1. Global News Monitoring

Identified key crisis indicators: disease outbreaks and economic risks
Discovered strategic opportunities in health partnerships and innovation sectors
Generated actionable recommendations for R&D investment allocation

2. Social Media Trend Analysis

Detected shift toward audience-focused content and emotional engagement
Identified brand visibility enhancement opportunities
Proposed targeted content campaigns leveraging emotional storytelling

3. Stock Market Intelligence

Analyzed positive growth patterns amid market volatility
Recommended diversified portfolio strategies and international market exploration
Suggested regular risk assessment adjustments

4. Industry Development Tracking

Highlighted digital transformation acceleration and supply chain adaptation needs
Identified skills gap challenges and production decline risks
Proposed employee training investments and supply chain technology enhancements

Output Documents Generated:

Response Strategies for Global Crises 2025
Stakeholder Contacts Spreadsheet for crisis coordination
Automated contingency plans with severity-based activation protocols

How I Used Runner H

My Prompting Strategy:

The key was creating a comprehensive prompt that encompassed multiple monitoring dimensions:

Monitor global news feeds, social media trends, market fluctuations,& industry
developments to instantly identify potential crises or opportunities.
Automatically draft response strategies, contact relevant stakeholders, create
contingency plans, monitor brand reputation in real-time, and execute 
predetermined actions based on severity levels without human intervention.

Implementation Steps:

Multi-source Integration: Connected news APIs, social media feeds, financial data streams, and industry reports
Intelligent Filtering: Applied AI-driven analysis to separate signal from noise
Automated Documentation: Generated structured response documents in Google Workspace
Stakeholder Mapping: Created dynamic contact lists based on crisis type and severity
Real-time Monitoring: Established continuous sentiment analysis frameworks

Use Case & Impact

Who Benefits:

Corporate Crisis Management Teams: Instant threat detection and response preparation
Investment Firms: Real-time market opportunity identification
Brand Managers: Continuous reputation monitoring and protection
Strategic Planners: Proactive risk assessment and contingency planning

Process Improvements:

Response Time: Reduced from hours/days to minutes
Coverage: 24/7 monitoring vs. business hours only
Consistency: Standardized response protocols vs. ad-hoc reactions
Cost Efficiency: Automated monitoring vs. dedicated human resources
Scalability: Simultaneous multi-dimensional monitoring vs. sequential analysis

Real-World Applications:

Financial institutions can instantly adjust portfolios based on emerging market signals
Healthcare organizations can prepare for disease outbreaks before they become widespread
Tech companies can capitalize on industry shifts as they develop
Consumer brands can address reputation threats before they escalate

This system transforms organizational resilience from reactive damage control to proactive opportunity capture, ensuring businesses stay ahead of both threats and market possibilities.

Social Love

Share your thoughts and experiences with automated crisis management in the comments below!

Also Check this same post out on Instagram