Inspiration

In George Orwell's 1984, surveillance was enforced by a totalitarian regime. In 2026, we surrender our digital footprint voluntarily across dozens of platforms, forums, and developer tools. Most users assume their exposure ends when they set their Instagram or Twitter to private. 

In reality, adversaries, recruiters, and OSINT doxxers correlate identity using abandoned 2018–2021 forum accounts with identical handles, unredacted contact numbers in hackathon presentation PDFs, exposed credentials in public GitHub dotfiles, and aggregator records on data brokers like Radaris.

We built ShadowTrace to turn modern neural search into a shield: discovering what the open web knows about you before an adversary or employer does.

What it does

ShadowTrace is an agentic reconnaissance and privacy defense copilot built specifically for the Surge Track ("Search Done Right"):

  • Target Reconnaissance: Enter an online handle, full name, or email. ShadowTrace dispatches parallel neural queries across Exa (semantic handle correlation and code leaks) and Tavily (deep context extraction for data brokers and public documents).

  • Calculated Exposure Gauge (0–100): Real-time risk scoring across 5 threat categories: Code Secrets & Repos, Data Brokers, Ghost Accounts, PII & Document Leaks, and Social OSINT.

  • Verifiable Receipts & Citations: Every single vulnerability card provides an exact source domain, publish date, search engine attribution, and direct quote snippet with a clickable link to verify the primary source with zero hallucinations.

  • 1-Click Legal Takedown Generator: For any exposed record, ShadowTrace auto-generates a legally binding GDPR Article 17 / CCPA "Right to be Forgotten" erasure letter pre-addressed to the target domain's Data Protection Officer (DPO), with one-click mail client launch (mailto:).

  • Remediation Roadmap & Offline Export: Provides an actionable defense checklist (Immediate, 24-Hour, Weekly) and one-click Markdown (.md) security audit report export.

How we built it

  • Framework: Next.js 16 (App Router) + React 19 + TypeScript.

  • Search & Neural Engine: Exa.ai (neural semantic search) + Tavily (deep content extraction).

  • Styling: Tailwind CSS + Custom Dark Cyberpunk Cyber-Security Theme.

  • Visual Radar Logo: Handcrafted SVG animated radar with concentric scanning rings and pulse gradients.

  • Zero Heavy Storage Waste: Cloud-agent architecture—no multi-gigabyte models required to download.

  • Demo Reliability: Built-in fallback demo engine ensuring judges can test full functionality instantly even without entering API keys.

Challenges we ran into

Filtering noise from open-web search queries while ensuring that every returned citation was click-testable and held up under examination. We solved this by designing a synthesis pipeline that pairs Exa's semantic intent matching with Tavily's deep raw text extract.

Accomplishments that we're proud of

  • Fulfilling the Surge Track criteria word-for-word: Tavily and Exa serve as the core product engine rather than an afterthought.

  • Clean 100% testable demo with zero hallucinations: all citations link to real sources.

  • Turning an audit into immediate action with the 1-Click GDPR/CCPA Takedown generator.

What's next for ShadowTrace

  • Continuous automated background monitoring with email alerts when new data leaks appear.

  • Direct API integration with broker opt-out networks (DeleteMe / Incogni style).

  • Browser extension to alert users whenever their email or handle appears on a site they are visiting.

Built With

Share this project:

Updates

Submission history