mcp-rag-agency-book-appointments

🗓️ AI-Powered Appointment Booking System

✨ Overview

This project is a cutting-edge AI-driven appointment booking system designed for agencies to schedule appointments, and manage queues in real-time efficiently. With AI Chat Bots, Retrieval-Augmented Generation (RAG), and MCP Client/Server integration, the system enables context-aware decision-making using external knowledge sources.

Built with Domain-Driven Development (DDD) principles, the system leverages Apache Kafka & ElasticSearch for appointment indexing, ensuring optimized search & performance across thousands of appointments! The architecture is microservices-ready, making it scalable and modular for enterprise adoption. Now powered by Qdrant Vector DB, it supports AI-driven semantic search and retrieval for enhanced user experiences.

🚀 Key Features

✅ Built-in IdentityServer with Asymmetric JWT Signing 🔐
✅ AI Chat Bots via PromptAPI 🤖
✅ Retrieval-augmented generation (RAG) – combines real-time knowledge retrieval with language generation for smarter, context-aware responses 🧠🔍
✅ Intelligent document parsing using LLM – extracts clean, structured, semantically meaningful text from messy HTML and scanned PDFs 📄🧹🤖
✅ Users can update the AI brain using RAG by scanning URLs & PDF documents on the fly – parsed content is embedded and stored in Qdrant for semantic search 📄🌐⚡
✅ RAG with Hangfire for document scan, parse & upload to Qdrant ⚙️
✅ MCP client/server ready ⚡
✅ Supports off days & max daily appointments 📅
✅ Real-time queue grid via API ⏳
✅ Domain-driven development (DDD) architecture 🏗️
✅ Event dispatcher for domain events 📨
✅ Apache Kafka & ElasticSearch for appointment indexing 📡
✅ Qdrant vector database for AI semantic search 🧠✨
✅ FluentValidation for validation logic ✅
✅ Swagger, LINQ, IoC, WebAPI 🛠️
✅ Automatic email template generation by AI LLM 📧✨
✅ API with Brain – users can type prompts in natural language 🧠📝
✅ Microservices-ready, modular & scalable 🏢🔄
✅ Cloud-ready (Azure/AWS) ☁️
✅ Refit-powered REST API clients 🔌
✅ Producer-consumer pattern with buffer cache for real-time insert, save & update 📤📥
✅ Next.js + React.js chatbot UI – real-time chat interface integrated with backend LLM API 💬⚛️ – supports theme changes
✅ Redis implementation for faster public key retrieval and chat store caching 🧰🔑

📜 Architecture Diagram

User → API Gateway → Appointment Service → Event Processing (Kafka) → Search Index (ElasticSearch)  
                  ↳ AI Decision Layer (RAG, MCP Client/Server, Qdrant Vector DB)  

🔄 User Flow

1️⃣ User registers on the platform 📝
2️⃣ Admin assigns "Agent" role to the user 👤✅
3️⃣ Admin registers an agency to the system 🏢
4️⃣ Agent adds agency users/customers (who will book appointments) 👥
5️⃣ Agent schedules an appointment for an agency user/customer 📅
6️⃣ AI automatically generates an appointment confirmation email template ✉️🤖
7️⃣ Appointment is indexed in Apache Kafka & ElasticSearch for real-time search 📡
8️⃣ User/customer gets notified with details via AI-enhanced email template 🚀
9️⃣ User interacts with AI freely via API with Brain – type any prompt, get smart AI responses 🧠💬
🔟 Qdrant Vector DB enhances search accuracy with AI-powered similarity matching 🔍💡
1️⃣1️⃣ Microservices-ready architecture ensures efficient scaling across multiple agencies 🏢⚙️

This ensures a streamlined booking experience, allowing agencies to manage appointments efficiently with real-time indexing, AI-generated email templates, and AI-driven semantic search with Qdrant!

🧰 Tech Stack

🟦 .NET 9 – modern, performant runtime for cloud-native applications
🛡️ IdentityServer – secure authentication and token issuance
📅 Hangfire – background job scheduling for asynchronous workflows
📡 Apache Kafka – distributed event streaming platform
🔍 ElasticSearch – high-speed, full-text search for appointment indexing
🧠 Qdrant – vector DB for semantic AI search
🧾 PromptAPI – LLM-based AI chatbot integration
🔌 Refit – declarative REST API clients with interface-based contracts
✅ FluentValidation – fluent rules for robust input validation
🧪 Swagger / OpenAPI – API documentation and test interface
☁️ Azure / AWS Ready – cloud-native infrastructure compatible
📜 Marten DB (PostgreSQL) – event sourcing and document database
🐘 PostgreSQL – backing store for MartenDB event sourcing and ElasticSearch sync
🗄️ MS SQL Server – primary application database for transactional data
🧱 producer/consumer repository pattern – buffer-backed async layer for write-heavy workloads
⚛️ Next.js / React.js – fast, modern frontend framework for interactive chatbot UI
🧠 Redis – in-memory cache for fast public key access and general-purpose caching

🛡️ Security & Access Control

⚠️ Strict access policies & authentication layers
🔐 JWT-based authentication
🔄 Audit logs for booking activities

📬 Contributing

We welcome new features, bug fixes, and performance improvements. 🚀
Feel free to submit pull requests or open issues!

⚡ Future Enhancements

🔮 AI-driven appointment recommendations
📢 Automated notifications for schedule changes
📡 Machine Learning for capacity prediction

---

📜 License - Apache License 2.0 (TL;DR)

This project follows the Apache License 2.0, which means:

• ✅ You can use, modify, and distribute the code freely.
• ✅ You must include the original license when distributing.
• ✅ You must include the NOTICE file if one is provided.
• ✅ You can use this in personal & commercial projects.
• ✅ No warranties – use at your own risk! 🚀

For full details, check the Apache License 2.0.

---

💡 This system isn't just another booking tool—it’s an intelligent, scalable AI-powered solution.
Let’s reshape the future of scheduling with AI, event-driven processing, scalable microservices, and AI-powered search with Qdrant Vector DB! 🚀🔥

---

This project is based on my other project: https://github.com/myonathanlinkedin/productinfo-mcp-rag

---

📸 Screenshots of the chatbots