PLC Demo Ordering System

The TAMU Physics Lab Center is responsible for setting up labs used in the physics curriculum, building physics demonstrations, and for delivering physics demonstrations to physics professors. We currently have a website with all the physics demonstrations that can be ordered, but the current issue with the PLC website is that unless you already have a good understanding of what demonstrations we have in our storeroom there are a lot of demos that never get ordered because the categorization of demonstrations on our website is very broad. In order to help professors find the best demonstration for their lecture I am developing a CrewAI-powered multi-agent system that automates the end-to-end process of finding the perfect demonstrations for a lecture and ordering a lecture based on a professor's syllabus. The diagram below explains the ideal flow of the final working system.

Flowchart of Agents in Final Working Product

This project includes a React web UI with mock API and CrewAI agent templates. The goal is to integrate real crew invocations and connect to the live PLC website.

Project Structure

plc-multiagent-demo/
├── frontend/       # React + Vite web UI (mock)
├── backend/        # FastAPI mock API
├── agents/         # CrewAI agent templates (plc_agents)
└── README.md

Quick Start

1. Backend (Mock API)

cd backend
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

2. Frontend

cd frontend
npm install
npm run dev

Open http://localhost:5173

3. CrewAI Agents (Templates)

cd agents
pip install -e .
# Set OPENAI_API_KEY or other LLM keys in .env
python -m plc_agents.main profile "PHYS 218" "mechanics,projectile motion"

Flow

1. Profile – Professor enters course and topics.
2. Search – System finds relevant demos; if none, offers Finding Chatbot.
3. Finding Chatbot – Conversational search when initial results are empty.
4. Help Chatbot – Extra details, videos, consultation scheduling.
5. Order – Place order and add items to demo calendar.

Agent Crews

Crew	Diagram Section	Purpose
ProfileCrew	Green	Create professor profile from course + topics
FindingCrew	Blue	Search PLC database for relevant demos
ChatbotCrew	Yellow	Conversational demo discovery (NLP search)
OrderingCrew	Red	Create order, add to calendar
HelpCrew	Purple	Demo details, videos, consultation

Future Integration

• Replace mock API logic with CrewAI crew invocations
• Connect tools to real PLC database and NLP embedding service
• Add YouTube API for help chatbot
• Add consultation booking (e.g., Calendly)