Explore how Haystack components interact with LLMs, understand what makes agents different from traditional language models, and create your first working agent. And then, learn how to give your agent access to real-world tools like web search, transforming it from a passive responder into an active problem solver!

Welcome to the course!

Talking to LLMs with Components

What can a Haystack agent do?

Creating our first agent

Connecting tools to our agent

Meet Your Agent

Unlock the tools to make your agents smarter, more transparent, and easier to debug in this chapter. From tracing agent decisions with streaming callbacks to building reusable Haystack pipelines, you’ll transform your agents into with a range of capabilities, including running SQL queries. Get ready to customize, connect, and supercharge your Haystack pipeline!

Opening the black box: making agents transparent

Customizing our agent

Haystack pipelines: our agent's secret power

Querying an SQL database

Turning our pipeline into a tool for agents

Wrap-up and where to go next

Designing Robust Agents

Haystack Additional Resources

Learn to go beyond simple chat interfaces and build fully functional AI agents that can reason, make decisions, and use external tools to take action. Using Haystack, an open-source framework for LLM-based applications, you'll construct a healthcare agent capable of searching the web and querying structured patient data. Along the way, you’ll master key agentic design principles, build reusable pipelines, and integrate components into an intelligent system. By the end, you’ll not only understand how AI agents work — you’ll have built one from scratch.

<h2>Build AI agents that reason, retrieve, and act</h2>

AI agents can do more than generate text. They can plan, make decisions, and take actions based on the tools and data available to them. In this course, you’ll learn how to build intelligent AI agents using Haystack, an open-source framework for orchestrating large language models and external components.
You’ll create a healthcare agent that answers user questions by combining up-to-date online content with structured patient records.
<br></br>

<h2>Conntect your agents</h2>

You’ll start by using Haystack components to communicate with an LLM, then move on to creating your first agent. You’ll connect tools such as real-time web search and a SQL database query pipeline, giving the agent the ability to pull in both public and private information. You’ll also learn how to register Haystack pipelines as tools, making it easier to reuse structured workflows inside your agent.
<br></br>

<h2>Build and extend real-world workflows</h2>

By the end of the course, you’ll have a robust agent that combines LLM reasoning with external tools and structured pipelines. You’ll understand when to use agents, when to use pipelines, and how to build flexible systems that can use both. Whether you’re building assistants, retrieval-augmented generation systems, or domain-specific tools, you’ll be ready to create production-ready AI agents with Haystack.



Intermediate Python

Create a healthcare AI agent using Haystack, an open-source framework for orchestrating LLMs and external components.

Building AI Agents with Haystack

DevRel Engineer for Haystack at deepset

You built a real agent with Haystack!

Agents that can reason and act

What's next

Let's practice!

https://s3.amazonaws.com/assets.datacamp.com/production/course_43749/subtitles/course_43749_3661099431707cdd8cddb38f4f685d68.vtt

Designing Robust Agents - Wrap-up and where to go next

Wrap-up and where to go next

1. Wrap-up and where to go next

2. You built a real agent with Haystack!

3. Agents that can reason and act

4. Agents that can reason and act

5. Agents that can reason and act

6. Agents that can reason and act

7. What's next

8. Let's practice!

Create Your Free Account