🤖 AI Copilot RAG Platform
A robust AI-driven Copilot system combining multi-agent RAG architecture, domain-specific intelligence, and real-time data visualization, built using React, FastAPI, LangChain, and Gemini API.
🧠 System Architecture
The architecture powers intelligent routing between agents (Clinical, Food Security, WebUI) and retrieves answers from relevant PDF knowledge bases using FAISS and Gemini LLM.
🔍 Features Overview
📌 1. Prompt Routing to Domain-Specific Agents
Automatically classifies prompts and delegates them to specialized agents for focused and relevant responses.
📈 2. Real-Time Chart Visualizations
Generates interactive pie, line, and table charts based on Gemini's structured output. Supports data prompts across health, agriculture, economy, etc.
💾 3. Persistent Widgets with API Tracking
Save visualizations as reusable widgets. Each widget includes the chart, title, and API endpoint for future reference.
💬 4. Clean Chat Interface with Markdown Formatting
Displays Gemini responses with support for headers, bold text, bullet points, and structured formatting for clarity.
🧪 Extra Feature Snapshot
Minimalistic yet functional UI with responsive design and fast navigation between saved charts and chat history.
🛠️ Technologies Used
React.js, FastAPI, LangChain, Gemini API, FAISS, MQTT, Flask, Docker, Recharts, Librosa
🏁 Built During
Spring 2025 CMPE 280 Hackathon
San José State University — Web UI Design and Development
“Production-grade implementation with cutting-edge LLM design and polished UI/UX” – Faculty Review
- Samrudh Sivva
