Chandigarh University official sealSmart India Hackathon official logo

Team Phoenix

URL Defender is our Smart India Hackathon 2023 entry from Chandigarh University, in the Cyber Security & Blockchain domain. Five friends, one problem statement, and 36 hours on the clock.

How it happened

  1. The idea

    Five friends in one room

    We were sitting together like any other evening when someone said it out loud — the hackathon is coming, why don't we build something instead of just watching? By the end of that night we had a problem worth solving: people lose money to links that look completely normal.

  2. The plan

    Work split by strength, not by title

    Only one of us actually lived and breathed cyber security. So we planned around what each person was genuinely good at — Python and modelling, data collection and filtering, the interface, the deck, the demo script. Nobody was carrying the whole thing alone.

  3. Internal round

    Selected at Chandigarh University

    We presented the threat model and a working detector to the internal SIH jury at Chandigarh University and got picked in the Cyber Security & Blockchain domain.

  4. 36 hours

    The build itself

    SIH runs for 36 straight hours. Dataset cleaning in one corner, feature engineering in another, the UI coming together on a third laptop, and everything merged at 4 a.m. A team doesn't break when it works together — that sentence stopped being a poster and became the whole night.

  5. The verdict

    The judges tried to break it

    They pasted their own links, real and fake, and watched the verdict land instantly. What impressed them was the combination — the accuracy of the calls and the speed at which they came back.

About Smart India Hackathon

SIH is a 36-hour nationwide hackathon organised every year by the Government of India, where student teams solve problem statements submitted by ministries, departments and industry. It is the largest open innovation model in the country — and if you're a college student, you can enter the next edition.

Apply at sih.gov.in

Kept alive since

The project didn't stop at the demo table. Incremental updates since then:

  • ·Dataset expanded to 88,647 labelled URLs across multiple public corpora
  • ·Domain-reputation arbitration layer added on top of the lexical classifier
  • ·Look-alike detection for character-swap impersonation (amazon → amozan)
  • ·Explainability panel: per-feature contribution to every verdict
  • ·Interface reworked for phones — one-hand reachable, offline capable