What it does This repo is the open-source backbone of promptingguide.ai, a sprawling reference on prompt engineering, RAG, and AI agents. It collects techniques (chain-of-thought, ReAct, tree of thoughts), model-specific notes, and a prompt library into MDX-based documentation you can run locally with Next.js.

The interesting bit The project treats prompt engineering as a genuine discipline, not a dark art. It maps academic papers to practical patterns—like translating research on “self-consistency” into something you can actually paste into a ChatGPT window. The 3 million learners stat suggests the approach is working.

Key highlights

• Covers 13 languages; actively seeking translations
• Includes a 1-hour video lecture, slides, and executable notebook
• Tracks techniques across models: GPT-4, Gemini, Mistral, LLaMA, Phi-2, OLMo
• Sections on risks and misuses (adversarial prompting, biases, factuality)
• Citable as an academic reference with provided BibTeX

Caveats

• The repo itself is mostly documentation glue; the real content lives on the website
• Local setup requires pnpm and Nextra, which is fine but not zero-friction
• Several course promotions and sponsorship callouts mixed into the README

Verdict Worth bookmarking if you’re building with LLMs and tired of scattered blog posts. Skip it if you want code libraries or hands-on frameworks—this is a reading list with structure, not a toolkit.