Prompt Engineering Guide

1. Project Overview

Prompt Engineering Guide is an open-source educational resource that consolidates research papers, techniques, tools, and hands-on examples for effectively designing prompts to get reliable, high-quality outputs from large language models (LLMs).

2. Background & Positioning

Created and maintained by DAIR.AI (Democratizing Artificial Intelligence Research), the project was started to address a common problem: as LLMs like GPT-4, LLaMA, and Gemini became more capable, developers and researchers had no single, structured place to learn how to actually prompt them well. Techniques were scattered across academic papers, blog posts, and tweets, with no consistent framework connecting theory to practice.

The guide's core mission is to make prompt engineering knowledge accessible to everyone — from complete beginners to researchers — by curating the latest papers, distilling them into plain-language guides, and pairing them with runnable notebooks.

What sets it apart from similar "awesome list" style repositories is its instructional depth: rather than just linking to resources, it builds a progressive curriculum (basics → advanced techniques → applications → risks), maintains a companion website (promptingguide.ai) with interactive content, and is available in 13+ languages, making it one of the most widely referenced prompt engineering resources in the LLM ecosystem.

3. Feature Categories

📘 Core Guides (10+ articles) — Structured lessons covering LLM settings, prompt basics, prompt elements, and general prompting tips. Examples: Prompting Introduction, Basics of Prompting, Prompt Elements, General Tips for Designing Prompts, Examples of Prompts.
🧠 Advanced Techniques (15+ techniques) — In-depth coverage of established and cutting-edge prompting methods grounded in published research. Examples: Zero-shot Prompting, Few-shot Prompting, Chain-of-Thought Prompting, Tree of Thoughts, ReAct Prompting, Retrieval Augmented Generation (RAG).
🛠️ Applications — Guides that translate techniques into practical use cases for building real products. Examples: Function Calling, Generating Data, Generating Code, Tackling Generated Datasets Diversity, Prompting for Graph Generation.
⚠️ Risks & Reliability — Guidance on the failure modes of prompting and how to mitigate them. Examples: Adversarial Prompting, Factuality, Biases.
🤖 Model-Specific Guides — Dedicated pages explaining prompting quirks and best practices for individual model families. Examples: ChatGPT, GPT-4, LLaMA, Gemini, Mixtral.
📚 Prompt Hub — A categorized library of ready-to-use prompt templates. Examples: Classification, Coding, Creative Writing, Evaluation, Information Extraction, Math, Reasoning, Summarization.
💻 Notebooks — Executable Jupyter notebooks that let learners run techniques against live models instead of just reading about them.

4. Key Highlights

Research-grounded content — Every technique is tied back to its source paper, so readers understand not just how to use a method but why it works.
Progressive learning path — Content is organized from fundamentals to advanced research topics, suitable for both newcomers and experienced practitioners.
Multilingual accessibility — Actively maintained translations in 13+ languages, driven by community contributors, extend reach well beyond English-speaking developers.
Companion interactive website — promptingguide.ai offers a searchable, navigable version of the guide with additional interactive elements not present in the raw Markdown/MDX source.
Executable notebooks — Pairing narrative guides with runnable code closes the gap between reading about a technique and applying it.
Massive, active community — With tens of thousands of GitHub stars and millions of learners reached, the project benefits from continuous community review and contribution, keeping it current with a fast-moving field.

5. Use Cases by Role

General developers — Learn how to structure prompts for chatbots, coding assistants, and content-generation features before wiring an LLM into an application.
Data & research scientists — Use the research-linked technique pages as a fast on-ramp to the prompting literature, and the notebooks to prototype experiments quickly.
Project managers / product owners — Skim the applications and risks sections to understand what's realistically achievable with current LLMs and where reliability caveats apply, informing scoping decisions.

6. Getting Started

Find what you need — Browse the guide directly on the website or search the repository's /guides and /pages directories for a specific technique:

open https://www.promptingguide.ai

Use it in your own workflow — Clone the repository to read the raw Markdown/MDX content offline or reference it from your own documentation:

git clone https://github.com/dair-ai/Prompt-Engineering-Guide.git

Contribute — Fork the repo, add or improve a guide, notebook, or translation, then open a pull request:

git checkout -b my-contribution
# edit files under pages/ or notebooks/
git commit -am "Add: <short description>"
git push origin my-contribution

7. Project Structure

Prompt-Engineering-Guide/
├── pages/          # Web content for the promptingguide.ai site (MDX)
│   ├── techniques/ # Prompting technique pages (few-shot, CoT, ReAct, RAG, ...)
│   ├── applications/ # Applied use-case guides
│   ├── risks/       # Adversarial prompting, factuality, bias
│   └── models/       # Model-specific guides (GPT-4, LLaMA, Gemini, ...)
├── notebooks/       # Executable Jupyter notebooks
├── lecture/         # Slides and materials for the companion video lecture
├── img/             # Diagrams and illustrations used across the guide
└── README.md        # Project overview and navigation

The pages/ directory is the source of truth for the live website; most substantive contributions (new techniques, corrections) happen there.

8. Related Ecosystem

Upstream dependencies: The guide references and builds on outputs from major LLM providers and research labs (OpenAI GPT models, Meta LLaMA, Google Gemini, Mistral) — it documents how to prompt these systems rather than depending on them as software libraries.
Companion resources from DAIR.AI: The DAIR.AI Academy offers structured, self-paced courses that expand on the free guide content; a companion video lecture and notebook provide a condensed walkthrough.
Complementary projects: Works well alongside LLM application frameworks (e.g., LangChain, LlamaIndex) and evaluation tools — the guide covers what to write in a prompt, while those frameworks handle how prompts are orchestrated in an application.

9. License

✅ Use the guide's content for personal learning, internal training, or as a reference in commercial products (MIT License).
✅ Copy, modify, and redistribute the content, including in derivative works, provided the original copyright and license notice are retained.
❌ Do not present the content as your own original work without attribution.
ℹ️ The MIT License applies to the repository's text and code; third-party assets (e.g., logos of model providers referenced in guides) may carry their own usage terms.

10. FAQ

Q: Is this guide affiliated with OpenAI, Google, or Meta?
A: No. It is an independent, community-driven educational resource maintained by DAIR.AI that documents how to prompt models from various providers.

Q: Do I need to run the notebooks to benefit from the guide?
A: No — the Markdown/MDX guides are self-contained and readable on their own. Notebooks are provided for readers who want hands-on practice.

Q: How current is the content given how fast LLMs evolve?
A: The repository is actively maintained with frequent pull requests adding new techniques and model-specific guidance; check the commit history or the website for the latest updates.

Q: Can I translate the guide into my language?
A: Yes — translations are community-contributed. Check open issues/PRs for existing translation efforts before starting a new one to avoid duplicate work.

Q: Where can I ask questions or discuss techniques with other users?
A: Via the project's Discord community and GitHub Discussions/Issues linked from the README.

11. Quick Links

Repository: https://github.com/dair-ai/Prompt-Engineering-Guide
Official docs / website: https://www.promptingguide.ai
Contributing guide: https://github.com/dair-ai/Prompt-Engineering-Guide/blob/main/CONTRIBUTING.md
Community: Discord and GitHub Discussions linked from the repository README

12. Summary

Prompt Engineering Guide is one of the most comprehensive and widely used open-source resources for learning how to write effective prompts for large language models, backed by research citations, runnable notebooks, and an active global community. It is valuable for anyone building with LLMs — from developers shipping their first chatbot feature to researchers tracking the latest prompting techniques — because it turns a fragmented, fast-moving body of knowledge into a single, structured, continuously updated reference.