Aigent MoneyPenny - Project Overview

BlakQube MoneyPenny - MVP Integration of Hackathon Services

Problem Statement

MoneyPenny solves the complexity of cross-chain DeFi trading by providing an intelligent trading console that eliminates manual chain monitoring, optimizes execution, and provides AI-powered market insights.

---

Three Hackathon Services Integration

1. Tavily (AI-Powered Web Search)

Module: src/lib/aigent/core/adapters/tavily.ts

Current Implementation:

• Research Agent (research-agent edge function): Searches real-time web data for token fundamentals, market news, protocol updates
• AI Trade Advisor (ai-trade-advisor edge function): Enriches trading recommendations with current market sentiment and external data
• Context Enhancement: Provides up-to-date information beyond blockchain data

Usage Flow:

User asks "What's the latest on ETH?"   
→ Tavily searches news/analysis   
→ AI synthesizes findings   
→ Returns actionable insights  

API Key: Configured via TAVILY_API_KEY secret

---

2. CopilotKit (AI Copilot Integration)

• Auto Populate Intent: Selects and Populates Intent card ready for execution

Planned Integration:

• Contextual Help: Real-time guidance during trade setup
• Workflow Assistance: Step-by-step onboarding for new users
  Would Enable: Intelligent UI components that proactively assist users with trading decisions

---

3. Redis (High-Performance Caching)

Module: src/lib/aigent/core/adapters/redis.ts

Current Implementation:

• Quote Stream Caching (quotes.ts module): Stores last 20 live quotes for instant retrieval
• Fill History Caching: Maintains recent trade executions for performance analytics
• Session Management: Tracks active trading sessions across chains
• Data Aggregation: Reduces API calls by caching frequently accessed market data
• Intent Form Auto-completion: AI suggests optimal trading parameters based on market conditions

Usage Flow:

SSE stream receives QUOTE   
→ Redis caches for 60s   
→ UI fetches from cache (sub-ms latency)   
→ Backend processes without re-fetching  

Configuration: Uses REDIS_URL for Upstash or local Redis instance

---

Technical Architecture Summary

User → MoneyPenny Console  
  ├─ Tavily: Real-time market research & news sentiment  
  ├─ Redis: Ultra-fast quote/fill caching (< 1ms reads)  
  └─ CopilotKit: AI-guided trading workflows and actions  

Current MVP Status:

• ✅ Tavily: Fully integrated for research & trade advisor
• ✅ Redis: Active for quote/fill caching and session management
• ✅ CopilotKit: Auto populate Intent and open Intent card

Full CopilotKit Integration
Build Redis Analytics Panel
Improve Research UI

Initial aigent concept developed last week but complete rebuild with completely new codebase and functonality.