Click to view Demo

TD Hackathon – Fraud Detection

React + Electron desktop app (Vite, TypeScript).

Fraudly prevents merchant and phone fraud while it’s happening.

Modern fraud targets the uninformed, happens in seconds, and is getting harder to detect as fraudsters use AI to craft convincing deceptions. We built Fraudly because prevention should happen in real time, before money is lost.

• $14B worth of crypto was stolen through scams.
• Phone scams grew 1,400% year-over-year.
• Of links delivered via Discord, ~20% were phishing, ~20% malware, and ~6% linked to fraud.

• 84% of fraud can be reduced by a simple warning.

• https://www.chainalysis.com/blog/crypto-scams-2026/ (Chainalysis Team, 2026)
• https://www.bitdefender.com/en-us/blog/hotforsecurity/bitdefender-scamio-is-now-on-discord (Alina, 2024)
• https://www.elliptic.co/blog/the-state-of-crypto-scams-2025-keeping-our-industry-safe-with-blockchain-analytics (Elliptic, 2025)
• https://www.fico.com/blogs/survey-consumers-want-better-scam-prevention-banks (Debbie, 2024)

• Malone Lam (2024): stole $260M+ with one phone call.
• The Prince Group / Chen Zhi (2015–2024): $15B stolen through phone/text scams.
• https://www.justice.gov/opa/pr/chairman-prince-group-indicted-operating-cambodian-forced-labor-scam-compounds-engaged (DOJ, 2025)
• https://www.justice.gov/usao-dc/pr/guilty-plea-and-superseding-indictment-announced-social-engineering-scheme-stole-263 (DOJ, 2025)

• We help everyone by reducing overall fraud.
• We help businesses build trust with consumers.
• We help consumers navigate fraud safely and confidently.

Fraudly helps detect and prevent fraud as it happens by delivering clear, actionable warnings when high-risk behavior is detected. It is designed to:

• protect consumers during online scam attempts including fraud through phone calls, email, and private messaging
• help merchants reduce fraud losses,
• build trust between businesses and customers.

We built Fraudly as a React + Electron app using Vite for fast iteration. The product focuses on a real-time warning experience so users can act immediately when suspicious patterns appear. And yes, over 70% of our code was written with AI.

• Designing warnings that are fast and clear without overwhelming users.
• Handling scam scenarios that evolve quickly across channels (phone, links, social platforms).
• Prompt Iteration for Deep Web Scraping on Company Terms of Service, Policies, and or Fraud Prevention Protocols
• Building a reliable cross-platform desktop experience.

• Turning complex fraud signals into simple, immediate warnings.
• Building an end-to-end demo that shows prevention in the moment.
• Creating a product vision that supports both consumers and businesses.

• Prevention timing matters more than perfect post-incident analysis.
• Even simple warnings can materially reduce fraud impact.
• User trust depends on clarity, speed, and low-friction guidance.

• Expand detection coverage for merchant and phone scams.
• Improve model precision and personalization of warnings.
• Add more integrations and deeper real-time risk context.
• Run larger pilot demos with partners and user groups.

npm i

Web only (browser):

npm run dev

Electron (desktop window):

npm run electron:dev

Starts the Vite dev server and opens the app in Electron with hot reload.

Web:

npm run build

Electron (packaged app):

npm run electron:build

Output is in the release/ directory.