Vibe Vibe is an open-source tutorial on Vibe Coding, written in Chinese under the
title 《Vibe Vibe —— 人人都能学会的 AI 编程（Vibe Coding）指南》 (“an AI programming guide anyone can
learn”) and published from the Datawhale community’s datawhalechina organization on GitHub. It is
read online at vibevibe.cn and licensed CC BY-NC-SA 4.0. Its stated audience is learners with no
programming background — students, entrepreneurs, working professionals, and also traditional
programmers wanting to learn a new way of working — and its stated aim is to carry them from “I have an
idea” to “I have built a product”, so that anyone can become a builder.
The tutorial takes its framing from Andrej Karpathy’s notion of vibe coding, which it glosses as a move “from Coder to Commander”: programming changes from writing code to conversational creation, directing an AI in natural language. It describes itself as the first systematic open-source vibe coding tutorial in China, motivated by what it sees as the fragmented state of Chinese-language learning material at a time when vibe coding has become a global trend.
Contents
The tutorial is organised into four parts, each aimed at a different reader:
- Foundations (基础篇) is an introduction for everyone. Its five chapters move from why now is the best time to program — including a chapter setting out vibe coding and “spec coding”, and a comparison of tools such as Cursor, Windsurf, Bolt.new and Replit — through thinking like a product manager (minimum viable products, the questions of who the user is and what the pain point is), the craft of talking to an AI (prompt basics under the heading “Context is King”, user journey maps, writing a first PRD, and prioritising features as P0, P1 and P2), a hands-on build of a personal tool in three rounds, and finally recovering from broken code, publishing to the web, security awareness, and the limits of what vibe coding can and cannot do.
- Advanced (进阶篇) is a sixteen-chapter guide to delivering a complete product “from zero to launch” while avoiding the usual pitfalls, covering among other things PRDs and document-driven development, environment variables and security, databases, testing, Git, serverless deployment and CI/CD, domains, server operations, SEO and analytics, and iterating on user feedback. It is built around a Next.js, React, TypeScript, Tailwind CSS, shadcn/ui, Drizzle ORM and PostgreSQL stack.
- Practice (实践篇) collects projects grouped by audience — humanities and business students, science and engineering students, working professionals — and an advanced set covering deployment, databases and authentication, AI agent development with RAG and MCP, full-stack projects, and workflow tools.
- Curated articles (优质文章篇) gathers company engineering blogs, podcasts, research reports, newsletters and developer communities for continued learning.
The README recommends different starting points by background: complete beginners start with the first chapter, readers who have used ChatGPT but never built a project start with the chapter on thinking like a product manager, and those with programming experience skim the foundations before moving to the advanced part.
Publication
Besides the hosted site, the repository ships a Dockerfile and Docker Compose configuration for deploying the tutorial site privately on a local or intranet network. The project also announces a forthcoming online edition built around a cloud IDE with Node.js, Python and Docker environments and more than fifty preinstalled AI skills, so that learners can start without installing anything.