LiveHeyGen ships HyperFrames Studio for Mac and Linux

AI · All levels

AI Agents — Complete Guide

A complete, from-first-principles course for understanding, designing, building, securing, evaluating, observing and deploying AI agents.

10 chapters30 pages

Complete curriculum

Choose a chapter

Every chapter and subtopic has its own detailed, indexable URL.

1What are AI agents?Understand agents, autonomy, goals, loops, and when an agent is actually useful.Open chapter

2 lessons

The agent loop Agent vs assistant

2Agent architectureLearn the components that turn a model call into a controllable agent runtime.Open chapter

2 lessons

The orchestrator Agent state machines

3Tool callingUnderstand typed tools, schemas, execution boundaries, validation, and safe side effects.Open chapter

2 lessons

Tool schemas Safe side effects

4MemoryLearn context, session state, durable memory, summarization, and retrieval boundaries.Open chapter

2 lessons

Context vs memory Long-running tasks

5RAG and knowledgeLearn retrieval-augmented generation, indexing, retrieval, ranking, grounding, and freshness.Open chapter

2 lessons

The retrieval pipeline Grounding and citations

6PermissionsDesign least-privilege agent capabilities and keep authorization outside the model.Open chapter

2 lessons

Least privilege Human approval gates

7Agent securityProtect agents from prompt injection, data leakage, unsafe tools, and untrusted content.Open chapter

2 lessons

Prompt injection Data exposure

8EvaluationMeasure agent quality with datasets, traces, assertions, graders, and regression tests.Open chapter

2 lessons

Evaluation datasets Trace graders

9ObservabilityTrace agent runs end-to-end and connect quality, latency, errors, and cost.Open chapter

2 lessons

Tracing agent runs Cost and latency

10Production deploymentMove an agent from prototype to reliable production software with budgets, retries, approvals, and rollback.Open chapter

2 lessons

Retries and idempotency Production release checklist