Product Catalog Chatbot with AI-Enhanced Management System

An intelligent product catalog management platform powered by Docker Model Runner, AI Agents, and Event-Driven Architecture. This system combines conversational AI, intelligent agents, and real-time processing for comprehensive catalog management.

🎯 System Overview

This is a complete AI-enhanced catalog management system featuring:

🤖 Core AI Components

Chatbot Interface - Natural language product queries and conversations
AI Agent Service - Automated vendor evaluation, market research, and customer matching
MCP Gateway - Model Context Protocol for AI tool orchestration
Model Runner Integration - Local AI model execution with Llama 3.2

🏗️ Complete Architecture

┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐
│   Frontend      │  │  Agent Portal   │  │  Chatbot UI     │
│   Port: 5173    │  │   Port: 3001    │  │   Port: 5174    │
└─────────────────┘  └─────────────────┘  └─────────────────┘
        │                     │                     │
        └──────────┬──────────┴─────────┬───────────┘
                   │                    │
        ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐
        │   Backend API   │  │ Agent Service   │  │ Chatbot API     │
        │   Port: 3000    │  │  Port: 7777     │  │  Port: 8082     │
        └─────────────────┘  └─────────────────┘  └─────────────────┘
                   │                    │                    │
                   └──────────┬─────────┴─────────┬──────────┘
                              │                   │
                    ┌─────────────────┐  ┌─────────────────┐
                    │  MCP Gateway    │  │  Model Runner   │
                    │  Port: 8811     │  │  (Local AI)     │
                    └─────────────────┘  └─────────────────┘
                              │
              ┌───────────────┼───────────────┐
              │               │               │
    ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
    │   PostgreSQL    │ │     MongoDB     │ │     Kafka       │
    │   Port: 5432    │ │   Port: 27017   │ │  Port: 9092     │
    │ (Products DB)   │ │ (Agent History) │ │ (Event Stream)  │
    └─────────────────┘ └─────────────────┘ └─────────────────┘

🚀 Quick Start

Prerequisites

Docker Desktop with Model Runner enabled
At least 8GB RAM (4GB+ for AI models)
Docker Compose v2.0+

1. Pull Required Models

# Pull the AI model for chatbot and agents
docker model pull ai/llama3.2:1B-Q8_0

2. Start All Services

# Clone the repository
git clone https://github.com/ajeetraina/catalog-service-node-chatbot.git
cd catalog-service-node-chatbot

# Start the complete system
docker compose up -d --build

3. Access the Applications

Service	URL	Description
🤖 Chatbot Interface	http://localhost:5174	Main chatbot for product queries
🏠 Main Frontend	http://localhost:5173	Product catalog management
🔧 Agent Portal	http://localhost:3001	AI agent management interface
📊 Kafka UI	http://localhost:8080	Event streaming monitoring
🗄️ pgAdmin	http://localhost:5050	Database administration

💬 Chatbot Capabilities

Note: After running sh add-products.sh, your chatbot will have access to 50+ sample products for testing these capabilities.

Natural Language Product Search

User: "Show me all electronics under $500"
Chatbot: "I found 8 electronics products under $500:

📱 iPhone SE - $399.00
   Compact smartphone with A15 Bionic chip

🎧 Sony WH-1000XM5 - $399.00  
   Industry-leading noise canceling headphones

⌚ Apple Watch - $249.00
   Advanced health and fitness tracking

[Additional products...]

Would you like details about any specific product?"

Intelligent Category Browsing

User: "What's popular in home electronics?"
Chatbot: "Based on our catalog data:

🏠 Smart Home Category (12 products):
• Smart speakers: 4 products ($49-199)
• Security cameras: 3 products ($89-299) 
• Smart thermostats: 2 products ($149-249)
• Home automation hubs: 3 products ($99-349)

The Amazon Echo Dot and Ring Doorbell are our top sellers.
Which type of smart home device interests you?"

Catalog Analytics

User: "Give me a summary of the catalog"
Chatbot: "Here's your catalog overview:

📊 **Catalog Statistics:**
• Total Products: 127
• Active Categories: 8
• Price Range: $9.99 - $2,499.00
• Average Price: $186.45
• New This Week: 5 products

🔥 **Top Categories:**
1. Electronics (34 products)
2. Clothing (28 products)  
3. Home & Garden (21 products)

Is there a specific category you'd like to explore?"

🤖 AI Agent System

The system includes intelligent agents powered by Model Runner:

🔍 Vendor Intake Agent

Purpose: Evaluates new product submissions
Scoring: 0-100 AI-powered evaluation
Criteria: Product quality, market fit, pricing analysis
Integration: Kafka events trigger automatic evaluation

📈 Market Research Agent

Purpose: Automated competitor analysis
Features: Price comparison, feature analysis, market positioning
Data Sources: Web scraping, API integrations via MCP Gateway
Output: Comprehensive market reports

🎯 Customer Match Agent

Purpose: Analyzes customer preferences and buying patterns
Intelligence: ML-based recommendation engine
Personalization: Tailored product suggestions
History: Stored in MongoDB for continuous learning

📋 Catalog Management Agent

Purpose: Maintains and optimizes product catalog
Automation: Auto-categorization, price updates, inventory sync
Quality Control: Duplicate detection, data validation
Optimization: SEO improvements, description enhancement

⚙️ Model Runner Configuration

Supported Models

Model	Size	Performance	Use Case
`ai/llama3.2:1B-Q4_0`	1GB	Fast	Chatbot, basic agents
`ai/llama3.2:1B-Q8_0`	1.5GB	Balanced	Recommended
`ai/llama3.2:3B-Q4_0`	2GB	High Quality	Complex agent tasks

Environment Configuration

# Model Runner integration in compose.yaml
models:
  llama_model:
    model: ai/llama3.2:1B-Q8_0

# Services using Model Runner
chatbot-backend:
  models:
    llama_model:
      endpoint_var: MODEL_RUNNER_URL
      model_var: MODEL_RUNNER_MODEL

agent-service:
  models:
    llama_model:
      endpoint_var: MODEL_RUNNER_URL  
      model_var: MODEL_RUNNER_MODEL

mcp-gateway:
  models:
    llama_model:
      endpoint_var: MODEL_RUNNER_URL
      model_var: MODEL_RUNNER_MODEL

🔧 Core Services Deep Dive

Chatbot Backend (`chatbot-backend`)

Port: 8082
Purpose: Natural language processing for product queries
AI Integration: Direct Model Runner connection
Database: PostgreSQL for product data
Features: Intent recognition, smart search, conversation context

Agent Service (`agent-service`)

Port: 7777
Purpose: AI agents for automation and intelligence
Event Processing: Kafka-based event handling
Databases: PostgreSQL (catalog) + MongoDB (history)
Agents: Vendor evaluation, market research, customer matching

MCP Gateway (`mcp-gateway`)

Port: 8811
Purpose: AI tool orchestration and integration
Protocols: Server-Sent Events (SSE) transport
Tools: fetch, brave, resend, curl, mongodb
Role: Enables agents to access external data sources

Backend API (`backend`)

Port: 3000
Purpose: Core catalog management API
Database: PostgreSQL primary, MongoDB secondary
Integration: Agent service communication
Features: CRUD operations, vendor management, analytics

📊 Event-Driven Architecture

Kafka Integration

# Kafka configuration for real-time processing
kafka:
  image: apache/kafka:latest
  ports: ["9092:9092", "9093:9093"]

# Event flow examples:
# 1. New product → Agent evaluation → Catalog update
# 2. User query → Agent research → Enhanced response  
# 3. Price change → Market analysis → Competitive insights

Event Types

Product Events: Creation, updates, deletions
Agent Events: Evaluation results, research findings
User Events: Interactions, preferences, feedback
System Events: Health checks, performance metrics

🗄️ Database Architecture

PostgreSQL (Primary Database)

-- Core product catalog
CREATE TABLE products (
    id SERIAL PRIMARY KEY,
    name VARCHAR(255) NOT NULL,
    description TEXT,
    category VARCHAR(100),
    price DECIMAL(10, 2),
    vendor_id VARCHAR(100),
    ai_score INTEGER, -- AI evaluation score
    status VARCHAR(20) DEFAULT 'active',
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);

-- Vendor management
CREATE TABLE vendors (
    id VARCHAR(100) PRIMARY KEY,
    name VARCHAR(255),
    evaluation_score INTEGER,
    status VARCHAR(20)
);

MongoDB (Agent History)

// Agent interaction history
{
  agentType: "vendor_evaluation",
  productId: "12345",
  evaluation: {
    score: 85,
    reasoning: "High quality, competitive pricing",
    criteria: ["quality", "price", "market_fit"]
  },
  timestamp: "2025-01-15T10:30:00Z"
}

// Customer interaction patterns
{
  sessionId: "abc123",
  interactions: [
    { query: "electronics under $200", results: 5 },
    { action: "view_product", productId: "67890" }
  ],
  preferences: ["electronics", "budget_conscious"]
}

🛠️ API Endpoints

Chatbot API (`localhost:8082`)

# Natural language chat
POST /api/chat
{
  "message": "Show me laptops for programming"
}

# Product search with filters
GET /api/products/search?q=laptop&category=electronics&maxPrice=1500

# Get conversation context
GET /api/context/{sessionId}

Agent Service API (`localhost:7777`)

# Trigger vendor evaluation
POST /api/agents/evaluate-vendor
{
  "vendorId": "tech_corp_001",
  "products": ["product_123", "product_456"]
}

# Get market research report
GET /api/agents/market-research/{productId}

# Customer preference analysis
POST /api/agents/analyze-customer
{
  "customerId": "user_789",
  "interactionHistory": [...]
}

Main Backend API (`localhost:3000`)

# Product management
GET /api/products
POST /api/products
PUT /api/products/{id}
DELETE /api/products/{id}

# Vendor management  
GET /api/vendors
POST /api/vendors
GET /api/vendors/{id}/evaluation

# Analytics
GET /api/analytics/summary
GET /api/analytics/trends

🔍 Development & Testing

Local Development Setup

# Start infrastructure only
docker compose up postgres mongodb kafka -d

# Run services locally for development
cd chatbot-backend && npm run dev  # Port 8082
cd agent-service && npm run dev    # Port 7777  
cd backend && npm run dev          # Port 3000

Health Checks

# Check all service health
curl http://localhost:3000/health   # Backend
curl http://localhost:8082/health   # Chatbot
curl http://localhost:7777/health   # Agents
curl http://localhost:8811/health   # MCP Gateway

# Check Model Runner status
docker model ls
docker model info ai/llama3.2:1B-Q8_0

Testing the Chatbot

# Test natural language queries
curl -X POST http://localhost:8082/api/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "What are your most expensive electronics?"}'

# Test product search
curl "http://localhost:8082/api/products/search?q=smartphone&maxPrice=800"

# Test category lookup
curl http://localhost:8082/api/categories

📈 Monitoring & Observability

Service Monitoring

Kafka UI: http://localhost:8080 - Event stream monitoring
pgAdmin: http://localhost:5050 - Database monitoring
Docker Logs: docker compose logs [service-name]
Health Endpoints: Built-in health checks for all services

Performance Metrics

# Check resource usage
docker stats

# Monitor model performance
docker model stats ai/llama3.2:1B-Q8_0

# View service logs
docker compose logs -f chatbot-backend
docker compose logs -f agent-service

🚀 Adding Products & Data

Automated Product Import

# Use the provided import script
./add-products.sh

# Or use the Node.js automation script
npm run import-products

Testing with Sample Data

The system includes comprehensive sample data: - 50+ Products across multiple categories - Vendor Information with AI evaluations - Mock Market Data for agent testing - Customer Interaction Patterns for recommendation testing

🔒 Production Considerations

Security

Environment Variables: Use Docker secrets for production
Network Security: Configure proper firewall rules
Database Security: Enable PostgreSQL/MongoDB authentication
API Security: Implement rate limiting and authentication

Scaling

Horizontal Scaling: Multiple instances of each service
Database Sharding: Partition large datasets
Model Optimization: Use quantized models for performance
Caching: Redis for frequently accessed data

Deployment

# Production environment variables
environment:
  - NODE_ENV=production
  - MODEL_RUNNER_URL=${MODEL_RUNNER_URL}
  - POSTGRES_HOST=${POSTGRES_HOST}
  - KAFKA_BROKERS=${KAFKA_BROKERS}
  - MONGODB_URI=${MONGODB_URI}

🎯 Quick Start Checklist

[ ] Pull AI Model: docker model pull ai/llama3.2:1B-Q8_0
[ ] Start Services: docker compose up -d --build
[ ] Import Products: ./add-products.sh
[ ] Test Chatbot: Visit http://localhost:5174
[ ] Check Agents: Visit http://localhost:3001
[ ] Monitor System: Visit http://localhost:8080