SuperioAI

Conversational AI for on-chain analysis and instant cross-chain execution

SuperioAI

Created At

ETHOnline 2025

Project Description

Superio AI is an AI-powered DeFi intelligence platform that helps users interact with blockchain and cryptocurrencies through natural language conversation, combining multiple AI agents, blockchain integrations, and real-time data sources.

Core Capabilities Natural Language Interface Users chat with AI to execute DeFi actions:

Token Swaps: "swap 0.001 ETH to FET" generates signable transaction Chart Analysis: "analyze ETH chart" AI-powered technical analysis Transaction Lookup: "explain this transaction: 0x..." detailed on-chain insights Yield Discovery: "show me high-yield pools" interactive knowledge graph

Multi-Agent AI System Coordinated specialized agents handle different tasks:

Coordinator Agent - Routes and orchestrates requests Swap Agent - Token swap intent detection and execution Trading Agent - Chart analysis with Gemini Vision AI Blockscout Agent - On-chain blockchain data queries DeFi/Coin/FGI Agents - Market data, cryptocurrency info, and sentiment

AI-Powered Chart Analysis

Fetches TradingView charts via Chart-IMG API Analyzes with Google Gemini Vision 2.5 Generates BUY/SELL/HOLD recommendations Provides entry/exit points with risk management

Blockchain Transaction Execution

Users initiate swaps through chat AI prepares signable transaction UI Web3 wallet integration (MetaMask, WalletConnect) Executes swaps on smart contracts

On-Chain Analytics

Transaction details, gas analysis, and confirmations Address analytics with reputation scoring Token holdings and transaction history Multi-chain support (Sepolia, Mainnet)

MeTTa Knowledge Graph

Interactive visualization of DeFi ecosystem Nodes for pools, tokens, and blockchains Color-coded relationships and security indicators Force-directed layout for exploration

Technical Stack Frontend

Next.js 14, React 18, Tailwind CSS wagmi/viem for blockchain interactions RainbowKit for wallet connection PWA capabilities

Backend

Flask REST API Multi-agent orchestration system ASI1-Mini LLM for intent classification MongoDB for chat history Google Gemini for vision analysis

Blockchain

Ethereum Sepolia testnet Blockscout MCP Server for on-chain data Smart contracts for token swaps Web3 wallet integration

Key Features

Conversational Interface - Chat naturally with AI Multi-Agent System - Specialized AI agents for different tasks Chart Analysis - AI-powered technical analysis with recommendations Token Swaps - Execute swaps through chat On-Chain Insights - Detailed blockchain analytics and reputation DeFi Discovery - Interactive knowledge graph visualization Chat History - Persistent, wallet-indexed conversations Context Awareness - AI maintains conversation context

Real-World Use Cases

"Swap 0.1 ETH to USDC" AI prepares swap transaction, user signs with wallet "Should I buy Bitcoin?" AI analyzes market sentiment and provides recommendation "Analyze this address: 0x..." AI provides on-chain profile and reputation score "What's the best yield farm?" AI searches and displays DeFi opportunities with APY "Explain this transaction: 0x..." AI breaks down transaction details and gas analysis

Current Deployment

Backend: Heroku (https://superio-c0e1ce720dee.herokuapp.com) Frontend: Vercel Blockchain: Ethereum Sepolia testnet Database: MongoDB Atlas

Future Enhancements

Multi-chain support (Polygon, Base, Arbitrum) Advanced trading strategies (limit orders, stop-loss) Portfolio tracking across multiple wallets Social features (shared trades, leaderboards) Mobile app (React Native)

Built with love using Fetch.ai ASI1-Mini, Google Gemini, uAgents, and modern Web3 technologies

How it's Made

Superio AI is built as a full-stack, AI-powered platform that bridges conversational interfaces with blockchain execution. This document explains the technical architecture, data flows, and implementation details.

System Architecture

High-Level Flow User Input → Frontend (Next.js) → Backend (Flask) → AI Agents → Tools/APIs → Response → Frontend Display Component Layers Layer 1: Frontend (Next.js 14)

Framework: Next.js 14 with App Router Language: TypeScript UI: React 18 components with Tailwind CSS State Management: React Hooks (useState, useEffect) Blockchain: wagmi (Hooks) + viem (Utilities) Wallet Integration: RainbowKit Deployment: Vercel

Layer 2: Backend API (Flask)

Framework: Flask (Python) Database: MongoDB (via PyMongo) AI Integration: OpenAI-compatible ASI1-Mini Deployment: Heroku

Layer 3: AI Agents (Python)

Framework: Custom Python modules Agents: Coordinator, Swap, Trading, Blockscout, DeFi, Coin, FGI Communication: Function calls (not distributed yet)

Layer 4: External Services

Chart-IMG API: TradingView chart images Google Gemini: Vision analysis Blockscout MCP: On-chain data CoinGecko: Market data

Detailed Component Breakdown

2.1 Frontend Architecture Chat Interface Purpose: Main conversational UI Technologies: Next.js, TypeScript, React 18, Tailwind CSS How it works:

User types message, triggers onSendMessage Message added to local state POST to /api/chat with message and wallet address Display response with text content, chart images, swap UI, and tools metadata Save to local state and chat history

Key Features:

Auto-scrolling to latest message Loading indicators Wallet address as user_id Chat history persistence

Message Component Purpose: Render individual chat messages Technologies: React, Markdown rendering Rendering Order:

Chart images (if tool.chart_url present) Markdown content Swap/Send UI components Yield pools (if present) MeTTa knowledge graph (if present) Tools used section

Special Handling:

Chart images displayed prominently at top Recommendation badges (BUY/SELL/HOLD) Error handling for failed image loads

Wallet Integration Technologies: wagmi, viem, RainbowKit How it works:

User clicks "Connect Wallet" RainbowKit shows wallet options User approves connection Address becomes available globally Used as user_id for chat history

2.2 Backend Architecture Flask Server Technology: Flask (Python) Database: MongoDB Deployment: Heroku

background image mobile

Join the mailing list

Get the latest news and updates