Wallet Shield

chatbot
upload files
analysis results 1
analysis results 2

Inspiration

Lots of our friends have crypto wallets, but they rarely understand the potential risks, or sometimes, even many crypto-related concepts. Enter Wallet Shield: it applies an AI algorithm to detect fraud in students' crypto wallets. It generates an easy-to-understand report and provides a visual summary chart, helping users quickly identify risks. A key consideration is the privacy of users' financial data - we store data securely and immutably on blockchain. Additionally, the app features an interactive chatbot that teaches users essential crypto literacy and where users can ask questions, ensuring they stay both informed and secure in the ever-evolving crypto landscape.

What it does

Wallet Shield performs three critical functions:

Fraud Detection: Using AI algorithms, Wallet Shield scans wallet transaction data for anomalies or patterns indicative of fraud.

Risk Reporting: It generates a comprehensive report of flagged transactions, visually representing risks through charts and summaries for easy interpretation.

Crypto Literacy: The built-in chatbot educates users about blockchain, crypto wallets, and common risks, acting as a personal tutor for crypto concepts.

Together, these features make Wallet Shield a powerful tool for improving crypto literacy and protecting users against threats.

How we built it

AI Algorithm: Isolation Forest, Scikit-learn, NumPy, JSON Parsing, Flask, Flask-CORS

Blockchain: Web3.py, Solidity, Alchemy, MetaMask, RPC URL

Frontend: React, Chart.JS, CSS, React-Chartjs-2

Interactive Chatbot: Built on Gemini's API

Challenges we ran into

Blockchain Complexity: While blockchain provides excellent security, integrating smart contracts for data storage proved challenging. Deployment and validating the contract caused lots of issues. There was a lot of learning involved in utilizing Blockchain specific API's and utilizing tools like Alchemy, Meta Mask to create our admin wallets.

AI Accuracy: Tuning AI algorithms to minimize false positives was time-consuming but critical for fraud detection.

Accomplishments that we're proud of

Successfully integrating AI-powered fraud detection and generating user-friendly reports and building an interactive chatbot that demystifies crypto concepts and teaches users effectively.

What we learned

Through this journey, we learned deep insights into how blockchain can enhance data security, a better understanding of fraud detection techniques and how to fine-tune AI algorithms for real-world use cases, and experience in creating seamless interactions between backend APIs, AI models, and frontend components.