⚡ Klar

What developers actually get hired for.

Klar reads the developer job market in real time and tells you which skills companies are genuinely hiring for — straight from live job postings. No influencer hot takes. No "you must learn X" threads. No predictions. Just the data.

👉 Try Klar live — no setup

🐙 Repo · 🐛 Report Bug · ✨ Request Feature

---

Everyone tells you what to learn. Almost nobody shows you the evidence. Klar is the evidence.

---

Table of Contents

• The Why
• The Problem
• What Klar Does
• Built Different
• Under the Hood
• Tech Stack
• Run It Locally
• Roadmap
• Contributing
• The Person Behind Klar
• License

---

The Why

I taught myself to code from a tier-3 college with a slow connection and a tight budget. The hardest part was never the syntax — it was the noise. Every day, a new thread told me to learn a different language, framework, or "must-have" skill. Roadmaps contradicted each other. Influencers sold urgency. None of them showed me a single shred of proof.

So I asked a simple question: what are companies actually hiring for, right now?

Not opinions. Not a 12-month-old survey. Not someone's personal roadmap. Live, current, verifiable demand.

That question became Klar.

The Problem

Developers — especially students and early-career engineers without insider networks — make high-stakes learning decisions based on vibes:

• Job boards show you listings to apply to, not patterns across the market.
• "Roadmaps" and influencer content are opinion, often months or years behind reality.
• Annual developer surveys are self-reported and stale the day they ship.

The result: people spend months learning the wrong things while the market quietly moves on.

Klar fixes the information asymmetry. It turns thousands of live job postings into a clear, honest picture of demand — so your effort goes where the market actually is.

What Klar Does

• 📊 Skill Demand Rankings — the most-requested skills right now, ranked by real posting counts, filterable by role (Frontend, Backend, Full-stack, DevOps, Data, Mobile).
• 🔎 Skill Search — type any skill and instantly see how the market values it.
• 🧠 Rich Skill Drawer — for any skill: live demand, remote share, the skills it most often appears alongside (co-occurrence), and disclosed salary insights.
• 💰 Salary Insights (disclosed-only) — real, employer-disclosed pay — never predicted or estimated — with full transparency on how many postings actually disclosed it.
• 🏢 Who's Hiring — top companies ranked by active openings, the skills each one asks for most, and their remote share.
• 🎯 Skill-Gap Advisor — compare your skill set against live market demand and see exactly where your gaps are.
• ⭐ Watchlist — save the skills you're tracking (secured with real authentication).
• 🔄 Daily Auto-Refresh — the dataset rebuilds itself on a schedule, so what you see is current.

Every number on Klar traces back to a real, active job posting. Descriptive, never predictive.

Built Different

	Job boards	Roadmaps / influencers	Annual surveys	Klar
Source of truth	Listings to apply	Opinion	Self-reported	Live listings, analyzed
Tells you what's in demand	❌	⚠️ subjective	⚠️ stale	✅
Real-time	partial	❌	❌	✅ daily refresh
Hype / predictions	—	❌ heavy	—	✅ none — facts only
Salary honesty	hidden	❌	aggregated	✅ disclosed-only + transparency

Klar isn't another job board. It's a lens on the market that no listing site gives you.

Under the Hood

Klar ingests postings from two independent live sources, normalizes and de-duplicates them, extracts skills against a curated taxonomy, and serves fast aggregated insights through a cached API.

Live Job Sources                  Klar Engine                      Client
┌──────────────┐    ingest   ┌──────────────────────────┐   API  ┌─────────────┐
│   Adzuna     │ ──────────▶ │  Normalize roles          │ ─────▶ │   React     │
│              │             │  Extract skills (taxonomy)│        │   Tailwind  │
│  Google for  │             │  Cross-source dedupe      │        │   + Framer  │
│  Jobs (JSearch)│ ────────▶ │  (company + title + city) │        │   Motion    │
└──────────────┘             │  Aggregate → Cache (6h)   │        └─────────────┘
                             └──────────────────────────┘

Engineering highlights:

• 🧩 Cross-source de-duplication — the same role often appears on multiple platforms. A normalized company + title + city key collapses true duplicates while preserving genuinely distinct openings, so counts stay honest.
• 🛰️ Multi-source ingestion — Adzuna + Google for Jobs (via JSearch), each behind its own quota-aware, sequential fetch loop with backoff so free-tier limits are never blown.
• 🧠 Skill taxonomy extraction — a single shared taxonomy parses messy job descriptions into clean, comparable skills (and normalizes role titles into six canonical lenses).
• ⚡ Aggregation + caching layer — MongoDB aggregation pipelines power every chart, fronted by a 6-hour in-memory cache and a warm-up routine for snappy first loads.
• 🔐 Production posture — Clerk authentication, Helmet, CORS allow-listing, and per-route rate limiting out of the box.
• 🎨 A UI that respects the data — an OnePlus-inspired black-and-torch-red system, spring-physics motion, and tactile 3D-tilt cards that make raw numbers feel alive.

Tech Stack

Layer	Technology
Frontend	React, Vite, Tailwind CSS v4, Framer Motion
Backend	Node.js, Express
Database	MongoDB (Atlas) + Mongoose
Auth	Clerk
Data sources	Adzuna API, Google for Jobs (JSearch)
Infra	Render (web + API), cron-based scheduled ingestion

Run It Locally

You don't need to. Klar is live and always current: https://klar-dev.onrender.com

But if you want to contribute or fork it:

Local setup for contributors

Prerequisites: Node.js 18+, a MongoDB URI, and free API keys from Adzuna + RapidAPI (JSearch).

git clone https://github.com/shaiksohelll/Klar.git
cd Klar

Create Server/.env and Client/.env (see .env.example for the full list of variables), then:

cd Server && npm install && npm run dev
cd ../Client && npm install && npm run dev

Trigger one ingest to seed data (use your own secret), then open http://localhost:5173:

curl -X POST http://localhost:5000/api/ingest/adzuna -H "X-Ingest-Secret: your_secret"

⚠️ Never commit your .env files or real keys — keep .env in .gitignore.

Roadmap

• Live multi-source ingestion (Adzuna + Google for Jobs)
• Cross-source de-duplication
• Skill demand rankings + role lenses
• Rich skill drawer (demand, remote, co-occurrence)
• Who's Hiring leaderboard
• Skill-Gap Advisor
• Disclosed-only salary insights
• Skill velocity badges (🔥 rising / ▼ cooling)
• More legal sources (Jooble, Remotive, Himalayas)
• Salary by experience level & region
• CI pipeline + test coverage
• Public, documented API

Known Limitations

• Salary bands apply to disclosed INR roles only. The salary band facet filters on salaryRange.currency === "INR" and salaryDisclosed === true. Jobs denominated in USD, GBP, CAD, or AUD — or jobs where the salary was predicted rather than disclosed — are excluded from salary-band results. Cross-currency normalization is on the roadmap.

Contributing

Klar is built in the open. If this resonates with you:

• ⭐ Star the repo — it genuinely helps more people find honest market data.
• 🐛 Open an issue for bugs or ideas.
• 🔧 Have an improvement? New data sources, a better skill taxonomy, and UI polish are especially welcome — open an issue first so we can chat.

The Person Behind Klar

Klar was designed, built, and shipped solo by Sohel — @shaiksohelll.

I'm 21, and I graduated from a tier-3 engineering college. No elite network, no Silicon Valley internship, no mentor handing me a roadmap — just a laptop, a patchy internet connection, and a stubborn refusal to believe that where you start has to decide where you finish.

I built Klar because I needed it. As a self-taught developer, I was tired of guessing what to learn next while everyone shouted different answers. So instead of asking louder voices, I went to the source: the job market itself. Every aggregation pipeline, every dedupe edge case, every 2 a.m. "why is this returning an empty array" — I learned it by building it, one failure at a time.

If you're reading this from a small town, a tier-3 college, or a tight budget: this repo is proof that the gate was never as locked as they told you. Build the thing you wish existed. I did.

• 🐙 GitHub: @shaiksohelll

License

Klar doesn't carry an open-source license yet — for now, all rights reserved. Want to use or build on it? Open an issue or reach out and let's talk.

---

If Klar helped you see the market clearly, drop a ⭐ — it means the world to a solo builder.

Built with sheer determination, one commit at a time.