GuardFlow

An AI-powered multi-device fraud prevention system that protects users from phishing attacks, fake payment portals, scholarship scams, QR frauds, and other online financial threats before money is lost.

GuardFlow combines an Android application, Browser Extension, AI PC Backend, and Arduino-based hardware authentication into a single intelligent security ecosystem. By correlating events across multiple devices, GuardFlow can detect suspicious activities in real time and warn users before they complete a risky transaction.

---

Project Description

Android App

The Android application acts as the central dashboard.

Responsibilities:

• Show AI explanations
• Receive backend responses
• Receive NFC verification
• User interaction

Browser Extension

Purpose : Stop phishing before users submit information.

How it works:

• Detects current webpage
• Extracts URL
• Sends URL to backend
• Waits for AI analysis
• Displays warning popup

AI PC Backend

The backend is the brain of GuardFlow.

Responsibilities:

• Receive requests
• Analyze websites
• Calculate risk
• Generate explanations
• Return recommendations

Arduino Device

Purpose: Provide secure, contactless identity verification using NFC technology.

Responsibilities:

• Read NFC card UID
• Send UID to Android App
• Verify authorized users
• Trigger access decision
• Enable real-time authentication

Features

• Detects phishing and scam websites
• Live webpage analysis through Browser Extension
• AI-generated fraud explanations
• Deterministic risk scoring (0–100)
• Multilingual voice alerts
• Real-time Android notifications
• NFC-based physical verification for high-risk payments
• FastAPI backend with WebSocket communication
• Local event storage and session tracking

---

Tech Stack

Component	Technologies
Android App	Kotlin, Jetpack Compose, CameraX(YOLO)
Backend	Python, FastAPI, SQLAlchemy, SQLite
Browser Extension	JavaScript, Chrome Extension (Manifest V3)
AI	Qualcomm GenieX, Qwen, Sarvam AI, Bulbul v3
Hardware	Arduino UNO R4 WiFi, RC522 NFC, Servo, LEDs, Buzzer
Communication	REST APIs, WebSockets, JSON

---

Complete System Setup

Step 1 — Clone the Repository

git clone https://github.com/your-username/GuardFlow.git

cd GuardFlow

---

Step 2 — AI PC Backend

1. Navigate to the backend folder.

cd AI_PC_Backend

2. Create a virtual environment.

python -m venv .venv

3. Activate the virtual environment.

Windows

.\.venv\Scripts\activate

Linux / macOS

source .venv/bin/activate

4. Install the required dependencies.

pip install -r requirements.txt

5. Create a configuration file.

cp .env.example .env

Open .env and update the required values:

SARVAM_API_KEY=your_api_key
LLM_MODEL=Qwen/Qwen2.5-3B-Instruct
HOST=0.0.0.0
PORT=8000

6. Start the backend server.

uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload

Verify that the backend is running:

http://localhost:8000/docs

---

Step 3 — Browser Extension

1. Open Chrome and navigate to:

chrome://extensions

2. Enable Developer Mode.

3. Click Load unpacked.

4. Select the Browser_Extension folder.

5. Update the backend WebSocket URL if required.

Example:

const WS_URL = "ws://<AI_PC_IP>:8000/ws";

6. Reload the extension.

7. Open the extension popup and verify it shows:

Connected

---

Step 4 — Android App

1. Open the project in Android Studio.

File
    → Open
        → Android_App/

2. Allow Gradle Sync to complete.

3. Update the backend API URL.

Example:

const val BASE_URL = "http://<AI_PC_IP>:8000/api/v1/"

4. Build the project.

Build
    → Make Project

or

Ctrl + F9

5. Run the application.

Run
    → Run 'app'

or

Shift + F10

6. Grant all required permissions:

• Accessibility Service
• Camera
• Notifications
• Internet

7. Verify that the app opens successfully and connects to the backend.

---

Step 5 — Arduino Device

1. Open the Arduino project in Arduino App Lab.

File
    → Open
        → Arduino/

2. Install the required libraries.

Open:

Tools
    → Manage Libraries

Install:

• Servo
• MFRC522

3. Connect the hardware.

Component	Pin
Green LED	D2
Yellow LED	D3
Red LED	D4
Servo	D5
Buzzer	D6
RC522	SPI

4. Configure the backend IP address.

Example:

BACKEND_BASE_URL = "http://<AI_PC_IP>:8000"

5. Upload the firmware.

Sketch
    → Upload

or click the Upload (→) button.

6. Register the authorized NFC card using its UID.

Run the discovery sketch.

Open the Serial Monitor.

Tools
    → Serial Monitor

Tap the NFC card.

Example output:

Detected UID:
73 A2 91 C4

Copy the UID into:

Config.h

AUTHORIZED_UID = {0x73, 0xA2, 0x91, 0xC4};

Upload the final firmware again.

7. Verify the device.

• LEDs turn on correctly.
• Servo locks and unlocks.
• NFC card is detected.
• Device communicates with the backend.

Start the Complete System

Start each component in the following order:

1. Start GenieX (Local AI)
2. Start the AI PC Backend
3. Load the Browser Extension
4. Launch the Android App
5. Power the Arduino Device
6. Perform the demo

Following this sequence ensures that every component can communicate correctly with the others.

---

Complete Workflow

User Opens Website
        │
        ▼
Android detects event
        │
        ▼
Backend receives event
        │
        ▼
Browser Extension analyzes webpage
        │
        ▼
Signals stored in Database
        │
        ▼
Risk Engine calculates score
        │
        ▼
AI generates explanation
        │
        ▼
Translation + Voice Generation
        │
        ▼
Android displays warning
        │
        ▼
High Risk?
   │              │
  No             Yes
   │              │
Continue     Arduino requests
             NFC Verification

---

Future Scope

• Cloud-based threat intelligence
• Machine Learning-based behavioral analysis
• Support for multiple browsers
• Banking API integration
• Smartwatch notifications
• Enterprise dashboard
• Additional language support
• Federated learning for improved privacy

---

License

This project is developed for educational, research, and hackathon purposes.

Feel free to use, modify, and extend the project with proper attribution to the contributors.

Team

Name	Email
Ashika	ashikajain2401@gmail.com
Chhavi	goelchhavi090@gmail.com
Lavanya	lavanyagoyal1212@gmail.com
Piyush	piyushkumarb2510@gmail.com
Uday	udayverma112006@gmail.com