Prompt Engineering Guide
Prompt Engineering Guide
A complete, free and regularly updated compilation of prompt techniques.
A compilation of prompt techniques from basic to advanced such as few-shot, chain-of-thought and ReAct, with research papers.
Overview
Prompt Engineering Guide is a site compiling knowledge on prompt-writing techniques for large language models, developed by DAIR.AI. Its content ranges from an introduction (model settings, elements of a prompt, tips and examples) to techniques such as Chain-of-Thought, ReAct and Tree of Thoughts, plus sections on AI agents, applications, models and risks.
The site's source code is published on GitHub under the MIT license and, according to the site, the content has many translations. Besides the free material, the site promotes separate paid courses on prompting and AI agents.
For businesses, this is a reference when you want to understand the names and workings of prompt techniques, independent of any particular AI provider.
Highlights
Easy-to-follow introduction
Explains model parameters, prompt structure, components and many examples by type of task.
Rich catalogue of techniques
Compiles many techniques such as zero-shot, few-shot, Chain-of-Thought, ReAct and Tree of Thoughts.
Provider-independent
Has dedicated sections on several models such as Claude, GPT and Gemini to compare approaches.
Risks and misuse section
A dedicated section on risks and how models can be misused, helping businesses use AI more safely.
Extended resources
Includes lists of research papers, tools, notebooks and datasets for deeper study.
Open source
The GitHub repository uses the MIT license, and you can run the site on your own machine.
What it's for
Looking up terminology
When you meet concepts like Chain-of-Thought or ReAct in documents, open the site to read the explanation and examples.
Building internal curricula
Training teams use the introduction as a framework to write staff guidance materials.
Designing AI agents
Engineering teams consult the AI Agents section when planning multi-step automation.
Assessing risk
Read the Risks & Misuses section before releasing a chatbot to customers.
How to use it
- 1
Read the Introduction
Start with model parameters, the basics and the components of a prompt.
- 2
View the examples section
Read prompt examples by task type such as summarization, classification and question answering.
- 3
Learn the techniques
Go through each technique in Prompting Techniques, trying it again with your business's data.
- 4
Read the risks section
Learn about the risks and how to prevent them before real deployment.
- 5
Run a local copy (optional)
For technical users: you need Node 18 or later and pnpm; install dependencies per the README, then start with the command below and open http://localhost:3000/.
$pnpm dev
Video guides
Prompt Engineering Overview
Elvis Saravia
Prompt Engineering Tutorial – Master ChatGPT and LLM Responses
freeCodeCamp.org
Prompt Engineering Guide - From Beginner to Advanced
Matthew Berman
Real-world example
Hypothetical example: a small accounting services company wants to use AI to classify incoming customer emails. The operations team lead reads the introduction and the few-shot technique on Prompt Engineering Guide, then writes a prompt with a few sample emails already labeled, such as fee inquiry, document submission and complaint. Before applying it, she reads the risks section to understand why AI shouldn't be left to act on unfamiliar content in emails, and has staff review the classification results in the early stage.
Tips & notes
Read selectively
The site is very broad; start with the introduction, then read in depth only the techniques you really need.
Combine with official documentation
Use this site to understand concepts, then cross-check with the dedicated guides from Claude, OpenAI or Gemini.
Try with real examples right away
Each technique should be tested on business data to see whether it works.
Distinguish free from paid content
The documentation is free to read; the courses promoted on the site are separate paid products.
FAQ
Who makes this site?
The site lists copyright as belonging to DAIR.AI.
Does it cost anything?
Reading the documentation is free and the source code is under the MIT license; the courses introduced on the site are paid.
Is there a Vietnamese version?
The repository says the guide supports multiple languages but a Vietnamese version could not be verified; only English was seen on the homepage.
Do I need to know programming?
Not to read it; you only need technical knowledge if you want to run the site on your own machine.
Useful links
Want to apply this in your business?
2dm looks at your workflows, picks the right tools and sets them up with your team. The first call is free.
Same topic
2dm Prompt Library
Prompts for marketing, sales, operations and learning. Fill in a few fields, then copy or open directly in ChatGPT or Claude.
Tailor-made sales landing page prompt
A prompt for Claude Code or a coding agent: build a sales landing page with its own identity for each product, honest about figures, with a lead form, stopping at a draft.
Claude prompting guide
4.5
Anthropic's official guide: how to write clear requests, give examples, break down tasks and let Claude think step by step.