Prompy

Learning Hub

Learn to write better prompts

Videos, guides and courses we'd point a friend to, from your first prompt to agents. Pick a level or a topic, and save what works in your notebook as you go.

21 resources · 9 h 28 min of video and courses · free to read and watch

Step 1 · Beginner

Start here

How models work and the habits that fix most prompts. No experience needed.

Only beginner (8)

VideoBeginner1 h

Intro to large language models

Andrej Karpathy's one-hour talk for a general audience: what a language model is, how it is trained, why it makes things up, and where the technology is heading. Start here if you want to know what you are actually talking to before you learn how to talk to it.

Andrej Karpathy#llm-basics
AnthropicDocs

DocsBeginner

Prompt engineering overview

Anthropic's official guide, and the best first read. It goes from the highest-impact habits to the finer ones: be clear and direct, give examples, let the model think, structure input with XML tags, assign a role, and chain prompts. Each page is short and has before and after examples you can copy.

OpenAIDocs

DocsBeginner

Prompt engineering guide

OpenAI's guide to prompting GPT models: message roles and instructions, formatting with Markdown and XML, few-shot examples, giving the model context, and how prompting a reasoning model differs from prompting a GPT model. Read it alongside the Claude guide to see which advice holds across vendors.

GoogleDocs

DocsBeginner

Prompting strategies

Google's guide to prompting Gemini: clear instructions and constraints, response formats, few-shot examples, adding context, and breaking a task into smaller prompts. It ends with a useful section on how to change a prompt that isn't working.

AnthropicCourse

CourseBeginner

Interactive prompt engineering tutorial

Nine chapters of notebooks with exercises that are checked for you, from basic prompt structure to avoiding hallucinations and building complex prompts for legal, finance and coding tasks. Do this after the overview: you learn far more by fixing a broken prompt than by reading about one.

DeepLearning.AICourse

CourseBeginner1 h 30 min

ChatGPT prompt engineering for developers

A free short course with Andrew Ng and OpenAI's Isa Fulford, aimed at people who will call a model from code. Two principles (write clear instructions, give the model time to think), then worked examples of summarising, inferring, transforming and expanding text, and a small chatbot. Runs in the browser.

Learn PromptingCourse

CourseBeginner

Learn Prompting

A free, open-source course that assumes no technical background. It starts with what a prompt is, moves through basic and advanced techniques, and has one of the clearer introductions to prompt hacking and how to defend against it. Good if you prefer short lessons in order.

Learn Prompting#fundamentals#security

VideoBeginner2 h 11 min

How I use LLMs

A practical tour of everyday use: when to start a new chat, when to reach for a thinking model, search, deep research, files, voice, images and custom instructions. Less about wording, more about choosing the right tool and giving it the right context.

Andrej Karpathy#llm-basics#workflows

Step 2 · Intermediate

Go deeper

Techniques, testing your prompts, and keeping them safe from injection.

Only intermediate (9)

VideoIntermediate1 h 16 min

AI prompt engineering: a deep dive

A round-table with four of Anthropic's prompt engineers. They cover what makes someone good at prompting, why you should write to the model as you would brief a smart colleague, how to read its output for clues, and which old tricks no longer matter. Watch it once you have written a few prompts of your own.

DAIR.AIArticle

ArticleIntermediate

Prompt Engineering Guide

A community reference to look things up in rather than read end to end. Each technique (zero-shot and few-shot, chain-of-thought, self-consistency, ReAct, retrieval-augmented generation, prompt chaining) gets a short explanation, an example and a link to the paper behind it.

GoogleArticle

ArticleIntermediate

Prompt engineering whitepaper

Lee Boonstra's whitepaper from Google. Covers sampling settings (temperature, top-K, top-P) and the main techniques with worked examples, then closes with a list of best practices, including how to document your prompt attempts in a table. That last habit is what a prompt notebook is for.

OpenAIArticle

ArticleIntermediate

GPT-4.1 prompting guide

OpenAI's cookbook guide to prompting a model that follows instructions literally. Has a full agentic system prompt you can study, advice on tool descriptions, long context and where to place instructions, and a recommended prompt structure.

AnthropicDocs

DocsIntermediate

Define success criteria and build evaluations

How to decide what "good" means before you tune a prompt, then test for it: specific, measurable success criteria, building test cases that include edge cases, and grading by code, by a model or by a person. The step most people skip.

Anthropic#evals#claude
Hamel HusainArticle

ArticleIntermediate

Your AI product needs evals

Hamel Husain's case study of how a real product team got unstuck: unit tests for prompts, looking at your data every day, human and model review, and why a cheap evaluation loop is what lets you change prompts with confidence.

Hamel Husain#evals

VideoIntermediate3 h 31 min

Deep dive into LLMs like ChatGPT

A three-and-a-half-hour walk through the whole stack: tokenisation, pretraining, fine-tuning into an assistant, hallucinations, tool use and reasoning models. Long, but it explains why the prompting advice in the other guides works.

Andrej Karpathy#llm-basics
Simon WillisonArticle

ArticleIntermediate

Prompt injection: the series

Simon Willison coined the term and has tracked it since 2022. Read the earliest posts for the core problem (a model cannot reliably tell instructions from data), then the recent ones on what it means for agents with tools. Required reading before you connect a model to anything private.

Simon Willison#security
OWASPDocs

DocsIntermediate

LLM01: Prompt injection

The first entry of the OWASP Top 10 for LLM applications. A reference page: direct and indirect injection, example attack scenarios, and a checklist of mitigations such as least privilege, output validation and human approval for risky actions.

Step 3 · Advanced

Build with it

Agents, context engineering and connecting models to tools.

Only advanced (4)
AnthropicArticle

ArticleAdvanced

Building effective AI agents

Anthropic's field notes on what works in production. Separates workflows from agents, describes five patterns (prompt chaining, routing, parallelisation, orchestrator-workers, evaluator-optimizer) and argues for the simplest thing that works. The best single essay on the subject.

AnthropicArticle

ArticleAdvanced

Effective context engineering for AI agents

Why prompting grows into context engineering once a model runs in a loop: treating context as a limited budget, writing system prompts at the right altitude, just-in-time retrieval, and compaction, note-taking and sub-agents for long tasks.

AnthropicArticle

ArticleAdvanced

Writing effective tools for AI agents

A tool description is a prompt too. How to choose which tools to build, name and namespace them, return only useful context, and improve them by running evaluations and reading the transcripts. Useful if you connect a model to your own tools over MCP.

Model Context ProtocolDocs

DocsAdvanced

What is the Model Context Protocol?

The official introduction to MCP, the open standard that lets an AI app use outside tools and data. Read it to understand what happens when you connect Prompy to Claude, Cursor or another client and your prompts show up there.

Model Context Protocol#mcp#tools