FireWallet

Home Project Portfolio Github Repository Project Link

Overview

EEG Neural Wearable System Diagram
Figure 1: Home page

FireWallet is a next-generation crypto wallet system built on the Polygon blockchain that introduces a novel form of token obfuscation through programmable graph-based sub-wallets. Designed to provide a defense-in-depth security model, it fuses smart wallet partitioning algorithms, graph traversal heuristics, and a rich, real-time UI extension that streamlines crypto activity and wallet anonymity.

In conventional Web3 applications, wallet privacy and behavioral predictability remain critical vulnerabilities. Wallet addresses are trivially traceable and often reused, creating transparent behavioral graphs that attackers can analyze. FireWallet responds to this by modeling wallet relationships as a dynamic directed graph — abstracting transaction flow and enabling token partitioning across n layers of sub-wallets.

Role

Key Features

Graph-Based Wallet Obfuscation (Neo4j + PageRank)

Each wallet is a node, and token transfers form directional edges. FireWallet scores sub-wallets using a modified PageRank system, ensuring that frequently accessed nodes become less likely token targets, while rarely used nodes gain priority. This makes transfer patterns harder to model and trace. Eulerian path logic ensures full token partitioning with zero token loss across N wallets.

Why this model? Graph-based obfuscation mimics real-world laundering behaviors (e.g., shell wallets) and makes machine-learning analysis more difficult for attackers.

EEG Neural Wearable System Diagram
Figure 2: Graph Wallet visualization

Partitioning Algorithm Thought Process

I used a probabilistic Eulerian traversal approach where edge weights dynamically shift based on transaction volume and recency. A reverse-weighting mechanism deprioritizes frequently hit paths. The goal: ensure every wallet is visited once (Eulerian), while load-balancing the token amount across wallets to prevent obvious centralization or honeypot behavior.

EEG Neural Wearable System Diagram
Figure 3: Eulerian Graph being made code snippet

PostgreSQL + ElasticSearch Backend

Postgres manages actual transfer queues and user sub-wallet mappings. ElasticSearch allows instant full-text wallet indexing and metadata lookups. I used Redis to cache Alchemy queries, reducing request time by 80%.

Why this structure?: A traditional relational DB ensures transaction reliability. ElasticSearch adds speed. Redis enables ultra-fast lookups during graph resolution.

Chrome Extension UI (Next.js + ChakraUI)

The wallet interface provides token balances, ERC20/NFT stats, and real-time wallet graph snapshots. ChakraUI was used to quickly theme the project and maintain responsiveness. Next.js allowed pre-rendering of token price data for faster dashboard loads.

EEG Neural Wearable System Diagram
Figure 4: Combined Wallet Features

API Integration

Architecture and Code Design

The backend was broken into microservices: Graph Engine (Neo4j), REST API (Express), and Cache Layer (Redis). Communication was abstracted with service-based controllers, allowing hot-swaps between CoinGecko and Alchemy data without affecting API consumers.

The frontend was modularly composed with hooks for auth, token lookups, price feeds, and user history — all communicating with the Graph Engine via lightweight REST endpoints. JWTs were used for user state persistence.

Challenges & Constraints

Outcome

Future Work

Tools Used