FinGuard

Project by NEXORA

Inspiration

FinGuard was inspired by a simple observation: many financial scams do not succeed because security systems fail, but because people make rushed decisions under pressure. Urgent invoices, unfamiliar recipients, suspicious payment methods, and convincing payment requests can cause users to act before verifying.

We wanted to create a solution that protects not only the transaction, but also the decision made before the transaction.

What it does

FinGuard is an AI-powered financial decision-safety platform that helps users evaluate a payment before sending money.

Users can manually enter payment details or upload invoices, bills, payment requests, and screenshots. FinGuard uses AI to extract important information and detect risk signals such as:

Unusual urgency or pressure
Unfamiliar recipients
Suspicious payment instructions
Unusual payment channels
Potentially risky transaction patterns

It then generates a safety assessment and, when needed, activates a Safety Pause that guides users through verification steps before they proceed.

The goal is not to block legitimate payments, but to help users pause, verify, and make a more informed decision.

How we built it

FinGuard was built using React, TypeScript, Vite, Node.js, Express, Supabase, and Google Gemini.

The core workflow is:

User Input / Document
        ↓
Supabase Storage
        ↓
Document Extraction
        ↓
Gemini AI Analysis
        ↓
Risk Signal Detection
        ↓
Safety Score
        ↓
Safety Pause & Verification