Community / Learn
Free courses, tutorials, research discussions, and practical guides to help you learn about AI agents, at your own pace.
A free practical introduction to LLM agents, tool use, smolagents, LlamaIndex, LangGraph, and Agentic RAG. Work through exercises and a final agent project. Includes English and Simplified Chinese editions; Python and basic LLM knowledge are expected. Model inference or local compute for exercises can have separate costs.
Open resourceA free self-paced introduction to delegating structured work to agents. Learn to provide context, define boundaries and checkpoints, and review results. Designed for people who already use AI at work. A ChatGPT account is required to start the course; product access for hands-on practice can depend on the learner's plan.
Open resourceA free short course for team leads moving from individual AI use to shared human-agent work. Covers roles, shared objectives, delegation, and appropriate information access. No technical background is required. The course is marked beta; practical use of Claude products may require separate access.
Open resourceA public engineering primer on choosing between fixed workflows and autonomous agents. Explains routing, task decomposition, parallel work, iterative evaluation, and tool design. Useful before choosing a framework. The article dates from 2024; consult current documentation before copying API or tooling details.
Open resourceA free official learning track covering agent components, tools, orchestration, guardrails, and implementation examples. Start with the concepts, then follow code and SDK references. Reading requires no account; running OpenAI API examples requires API access and may incur usage charges.
Open resourceAn official open-source course introducing agent concepts, tools, planning, memory, retrieval, multi-agent systems, and safety through lessons and code samples. Includes Simplified Chinese and other translations. Materials are free; current Microsoft Foundry examples can require an Azure account and billable services.
Open resourceA free developer course spanning API fundamentals, prompting, tool calling, retrieval, MCP, and agent architectures. Includes exercises and quizzes. Intended for Python developers who can handle JSON. The teaching material is free; hands-on model calls need an Anthropic API key and can incur charges.
Open resourceA free focused course on connecting agents to external tools and data with MCP. Build Python servers and clients, then work with tools, resources, prompts, and the inspector. Python, async/await, and API familiarity are expected. Reading and local protocol exploration are free; model-backed integrations can incur API fees.
Open resourceA free hands-on course on MCP concepts, SDKs, end-to-end applications, and custom workflow servers. Includes projects and an optional completion pathway. English and Vietnamese material is present in the official repository. Public reading is free; hosted exercises and model integrations may require accounts or paid inference.
Open resourceA free short course on creating reusable skills for Claude Code, configuring their instructions, testing when they activate, and sharing them with a team. Useful after learning basic agent tools. The lessons are public; following along in Claude Code can require separately paid product or API access.
Open resourceA public engineering guide to managing the information an agent receives across long tasks. Covers context selection, tool interfaces, retrieval, compaction, and memory. Best after a basic agent tutorial. Reading is free; applying the techniques may involve paid model calls.
Open resourceA public guide to evaluating agents that act over multiple steps. Learn how to define tasks, trials, grading rules, traces, and outcomes, then combine evaluation methods and diagnose failures. Suited to developers who already have an agent prototype. Reading is free; running evaluations can incur model and infrastructure costs.
Open resource