$git clone https://github.com/Azure-Samples/azure-ai-travel-agents:star: To stay updated and get notified about changes, star this repo on GitHub!
| 1 | <div align="center"> |
| 2 | |
| 3 | <img src="./docs/ai-travel-agents-logo.png" alt="" align="center" height="128" /> |
| 4 | |
| 5 | <h1>Agents and MCP Orchestration <br>with Langchain.js, LlamaIndex.TS,<br>and Microsoft Agent Framework!</h1> |
| 6 | |
| 7 | [](https://discord.gg/NcwHpz6bRW) |
| 8 | [](https://aka.ms/foundry/forum) |
| 9 | [](https://techcommunity.microsoft.com/blog/AzureDevCommunityBlog/introducing-azure-ai-travel-agents-a-flagship-mcp-powered-sample-for-ai-travel-s/4416683) |
| 10 | <br> |
| 11 | [](https://codespaces.new/Azure-Samples/azure-ai-travel-agents?hide_repo_select=true&ref=main&quickstart=true) |
| 12 | [](https://github.com/Azure-Samples/azure-ai-travel-agents/actions) |
| 13 |  |
| 14 | [](https://www.typescriptlang.org) |
| 15 | [](https://www.java.com) |
| 16 | [](https://dotnet.microsoft.com) |
| 17 | [](https://www.python.org) |
| 18 | [](LICENSE.md) |
| 19 | |
| 20 | :star: To stay updated and get notified about changes, star this repo on GitHub! |
| 21 | |
| 22 | [](https://github.com/Azure-Samples/azure-ai-travel-agents)   |
| 23 | |
| 24 | |
| 25 | |
| 26 | [Overview](#overview) • [Architecture](#high-level-architecture) • [Features](#features) • [Preview locally FOR FREE](#preview-the-application-locally-for-free) • [Cost estimation](#cost-estimation) • [Join the Community](#join-the-community) |
| 27 | |
| 28 |  |
| 29 | |
| 30 | [](#preview-the-application-locally-for-free) |
| 31 | |
| 32 | </div> |
| 33 | |
| 34 | ## Overview |
| 35 | |
| 36 | The AI Travel Agents is a **modular reference application** that leverages multiple **AI agents** to enhance travel agency operations. The application demonstrates how **LangChain.js**, **LlamaIndex.TS**, and **Microsoft Agent Framework** can orchestrate **multiple AI agents** to assist employees in handling customer queries, providing destination recommendations, and planning itineraries. Multiple **MCP** (Model Context Protocol) servers, built with **Python, Node.js, Java and .NET**, are used to provide various tools and services to the agents, enabling them to work together seamlessly. |
| 37 | |
| 38 | | Agent Name | Purpose | |
| 39 | | -------------------------------- | ----------------------------------------------------------------------------------------------------------------------------- | |
| 40 | | **Customer Query Unders |