Multi-Agent AI System
Multi-Agent AI System
A multi-agent AI system that accepts input in PDF, JSON, or Email (text) format, classifies the format and intent, and routes it to the appropriate agent.
Project Overview
The Classifier Agent detects both format and intent (e.g., RFQ, Invoice, Complaint, Regulation) and routes inputs to relevant agents. Each agent handles domain-specific tasks such as CRM-style data extraction, anomaly detection, and intelligent structuring.
The Shared Memory System preserves cross-agent context, enabling thread-based conversation tracking and metadata management. The system supports business-ready workflows like invoice parsing, RFQ handling, and compliance document analysis.
Key Features
- Modular design
- Real-time classification
- Context-aware processing
- Scalable for enterprise use cases
Tech Stack
ReactTypeScriptTailwind CSSVitePDF.jsPython (backend)Vitest (testing)
Snapshots


