Atlas

Problem Statement

Atlas is a risk analysis system specifically designed for the Hedera Hashgraph ecosystem. Built on a cutting-edge multi-agent architecture, the system employs specialized AI agents that work collaboratively to provide real-time portfolio insights, risk assessments, and market intelligence.

Solution

Atlas is built on a multi-agent architecture that's the heart of what makes this project unique. Instead of a traditional monolithic approach, I created specialized AI agents that communicate through a custom A2A (Agent-to-Agent) message bus . This allows for:Asynchronous processing of complex portfolio analysisReal-time coordination between different system componentsScalable architecture where new agents can be added without disrupting existing onesFrontend Technologies & Integration Core Stack:Next.js 14 with App Router - Modern React framework for server-side rendering and optimal performanceTypeScript throughout - Type safety across the entire frontendTailwind CSS - Utility-first styling for rapid UI developmentRadix UI - Headless components for accessibility and customization Blockchain Integration (The Tricky Part):Wagmi + RainbowKit - For EVM wallet connections@hashgraph/sdk - Official Hedera SDKCustom EVM-to-Hedera converter - This was particularly hacky but necessary State Management:@tanstack/react-query - Server state caching and synchronizationCustom IndexedDB storage - For persistent chat historyReact Hook Form - Form state with validationBackend Multi-Agent System This is where things get really interesting. I built 7 specialized agents :Chat Agent - Orchestrates user commands and coordinates other agentsWallet Agent - Handles Hedera account balance queriesPortfolio Agent - Calculates risk scores and portfolio metricsScanner Agent - Performs deep address analysisGraph Agent - Generates Chart.js configurations for visualizationsLLM Agent - Processes natural language with OpenAI GPT-4XTrend Agent - Analyzes X (Twitter) sentiment and market trends