ThreatScope

💡 Inspiration

In an age where anyone can unknowingly download or share harmful files, we wanted to build a simple, AI-powered tool that helps users understand exactly why a file might be dangerous — without needing cybersecurity expertise.

We were inspired by:

The rise of malware disguised as innocent scripts or text files
The lack of easy-to-use, privacy-respecting tools for file analysis
The idea that cybersecurity should be explainable, not intimidating

Our goal was to build something both technical and educational — a platform that empowers people to detect and understand digital threats in real-time, with no server or install needed.

🛠️ How We Built It

We built ThreatScope entirely in the browser using:

React + TypeScript for the UI and logic
TailwindCSS for fast, responsive design
Web Crypto API to generate SHA-256 fingerprints of each uploaded file
Simulated Natural Language Model (NLM) logic to explain threats like:
- Keyloggers
- Phishing scripts
- Ransomware loaders
- Obfuscated payloads
A ledger array that mimics a blockchain to store and detect past threats

We also implemented:

Local file scanning and pattern detection (no files leave the browser)
Natural language explanations for each threat type
Threat severity badges and a live threat ledger display
Voice alerts (via Web Speech API) for high-risk detections
A “Built with Bolt” badge for full compliance and design polish

🚧 Challenges We Faced

Making everything run 100% client-side, including file reading, scanning, and hashing — without any backend or server
Designing a UI that could show technical threat info in a friendly, human way
Ensuring real-time detection and explanation without lag, even for large files
Simulating a blockchain ledger system without actual distributed nodes
Detecting and handling edge cases like:
- Mismatched file extensions
- Large binary files
- Obfuscated JavaScript
Ensuring the entire UI was responsive for both desktop and mobile

🧠 What We Learned

How to use the Web Crypto API for secure, fast hashing in-browser
How to simulate cybersecurity analysis using pattern matching and NLM-style explanations
How to balance technical detection with usability and accessibility
How to build fully offline, privacy-first applications with zero backend
The power of clear, human-readable explanations in security tools