AI agents, explained
Three pieces for anyone new to the term. No coding background assumed, no theory of large language models required — each one starts from a concrete question.
Read them in order. The first settles what an agent actually is; the next two answer the questions people ask right after.
| Piece | Answers |
|---|---|
| What is an AI agent | How it differs from a chatbot, why it gets things wrong, how to start |
| Agents vs ChatGPT | Same models, different reach — which to use when |
| Do you need to code | How far you get without it, where the ceiling is, what raises it |
| Where model credit comes from | Buy a plan, attach your own key, let it choose the model, list your own agent — four routes, and how to mix them |
| Model tiers, and how to pick one | A tier is a group, not a model; what each suits, and what to do when you can't tell |
After these
If you're going to use agents, the next step is picking a tool. Platform categories splits the field into three kinds and describes the trade-off in each.
If you're using Aiglade, start with the guides.