Build a working AI agent from scratch in Python. Learn the think-act-observe loop, call an LLM, give your agent a tool it can use, and handle a failed tool call. Framework-agnostic and beginner-friendly.
An AI agent is a language model that can take an action, look at the result, and keep going until a task is done, not just answer once and stop. That single idea powers every agent framework you have heard of, and you can build it yourself in plain Python without any of them. This free, hands-on micro course walks you from zero to a working agent in about half an hour.
You will learn the think-act-observe loop that sits at the heart of every agent, make your first call to a language model, give your agent a tool it can request, and assemble the complete agent loop into a small research helper you can run. You will also see what happens when a tool call fails and how a few lines of error handling let your agent recover instead of crashing. The code examples use the Anthropic SDK directly, so the patterns transfer cleanly to any provider or framework, and you practice the loop logic right in the lessons with an in-browser Python editor.
This course is for students and self-learners with a little Python who are brand new to agents and want to build one for their own work, like a study helper, a research assistant, or a small task automator. Like every FreeAcademy course it is 100% free with no signup required to begin, and finishing the lessons and final exam earns you a certificate of completion. When you are done, the LangChain and CrewAI courses are the natural next steps.
3 modules • 7 lessons
No. A little Python is enough. If you can read a function and a for loop, you can follow along. The agent you build is plain Python plus one call to a language model, with no framework to learn first.
You can complete every lesson and the final exam for free. The runnable in-browser exercises need no key. To run the live agent on your own machine you need an API key with a small amount of credit, where each short call costs a fraction of a cent, but that is optional for finishing the course.
No. You build the agent from scratch with the Anthropic SDK so you understand the loop underneath every framework. Once you finish, the LangChain and LangGraph course is the next step for depth, and the CrewAI course covers multiple agents working together.
A small research-helper agent that can answer questions and call a tool when it needs an exact fact, starting with a calculator tool. You then learn how to swap in your own tools to build an agent for a study task, a research chore, or something you repeat.
Yes. Like every FreeAcademy course it is 100% free with no signup required to begin, and finishing the lessons and passing the final exam earns you a free certificate of completion you can add to your resume or LinkedIn.
Already know basic Python? Build one real thing end to end: a small AI data app that loads a dataset, sends a slice of it to a language model, and turns the response into a written analysis. You will use pandas only as much as the app needs, call an LLM API directly (no agents, no frameworks), wire the pieces into a single script with real error handling, and ship it as a shareable Streamlit app you can run locally. A project-based intermediate course with a free certificate of completion.
Build autonomous AI agents with Python using LangChain and LangGraph. Learn tool calling, stateful workflows, RAG-powered agents, multi-agent systems, and production deployment. From the ReACT pattern to a full customer support agent capstone.
Learn to build multi-agent AI systems with CrewAI in about 45 minutes. Understand the crew/agent/task model, define specialized agents, create tasks, and build a real-world market research crew with analyst, researcher, and writer agents.
Build a private, offline knowledge base over your own notes and PDFs using a local AI model. A hands-on beginner micro course: run a model with Ollama, turn documents into embeddings, store them in a local Chroma database, and ask questions answered from your own files. No cloud APIs, no fees, your data stays on your machine. Want to deploy it as a web app instead? See the Full-Stack RAG course (/courses/fullstack-rag-nextjs-supabase-gemini).
Set up Hermes Agent, the open-source self-hosted AI assistant from Nous Research that writes its own reusable skills and keeps all your data on your own machine. A beginner-friendly micro course: understand what makes Hermes different, install it with a single command, run your first useful tasks, and stay safe with a practical permissions checklist. Prefer a chat-app interface instead? Try the OpenClaw micro course (/courses/micro-openclaw-ai-agent).
Learn the most practical prompt engineering techniques in just 30 minutes. Master the 4-part prompt formula, specificity, iteration, and 5 power techniques that work across all AI tools.