A curated list of awesome Vibe Coding resources.
Vibe Coding（氛围编程）是由 Andrej Karpathy 在 2025 年提出的新兴编程范式，强调通过自然语言与 AI 工具协作完成软件开发。
- Contents
- General Resources
- Core Concepts
- AI Coding Tools
- Learning Resources
- Open Source Models
- Research Papers
- Related Awesome Lists
- Contributing
- License
- Vibe Coding 101 - Free, systematic, practical AI programming tutorials covering Claude, Cursor, Gemini and more
- Vibe Coding 零基础 AI 编程课 - From zero to commercial developer course
- Khan Academy - Computer Programming
- Coursera - AI and Machine Learning
- edX - Computer Science
- FreeCodeCamp
- The Odin Project
- Anthropic Claude Documentation
- Anthropic Claude API
- GitHub Copilot Documentation
- Cursor Documentation
- OpenAI API Documentation
- Cursor - AI-powered code editor
- GitHub Copilot - AI pair programmer
- Claude Code - Official CLI tool for vibe coding
- Aider - AI pair programmer in your terminal
- Continue - Open-source autopilot for VS Code
- Bolt.new - AI-powered web development platform
- v0.dev - AI UI component generator
- Div-idy - First fully AI-generated website builder platform
- Void - Open-source AI code editor based on VS Code, emphasizing privacy and local control
- Codium - AI-powered test generation
- Tabnine - AI code completion
- CodeWhisperer - AWS AI coding companion
- Bloop - AI code search and understanding
- 3Blue1Brown - Math and AI visualizations
- Andrej Karpathy - Neural Networks: Zero to Hero
- StatQuest - Statistics and machine learning
- Traversy Media - Web development tutorials
- FreeCodeCamp - Programming tutorials
- The Coding Train - Creative coding
- Fireship - Quick tech tutorials
Vibe Coding is a programming paradigm introduced by Andrej Karpathy in February 2025, emphasizing collaboration with AI tools through natural language to complete software development.
"I just write code, the AI writes the code." - Andrej Karpathy
- AI Collaboration First: Developers become "Managers", AI assistants (like Claude, GPT-4) handle code implementation
- Natural Language Interaction: Describe requirements in human language, iterate through prompts
- Rapid Iteration: Focus on quick implementation of ideas and features, not syntax details
- Formal Definition: According to YuyaoGe's research, Vibe Coding can be viewed as a Constrained Markov Decision Process (CMDP) with multiple development patterns
- Intent-First: Start with "What do I want to implement" rather than "How to write code"
- Iterative: Continuously refine and improve through dialogue
- Low-Friction: Reduce cognitive burden, let AI handle tedious syntax and framework details
- Democratizing: Lower the barrier to programming, enabling more people to create software
📝 Vibe Coding Concepts - Detailed analysis of development patterns
According to research, Vibe Coding includes 5 main development patterns:
- Unconstrained Automation: Fully automated code generation
- Iterative Dialogue: Step-by-step refinement through conversation
- Planning-Driven: Generate plans first, then implement
- Test-Driven: Write tests first, then generate code
- Context-Enhanced: Leverage project context for better results
📝 Best Practices Guide - Comprehensive best practices and case studies
- Be Specific: Clearly describe what you want to implement
- Provide Context: Explain project background, tech stack, constraints
- Iterate and Refine: Gradually improve requirements through multiple rounds of dialogue
- Request Explanations: Ask AI to explain generated code for deeper understanding
- Verify and Test: Always test AI-generated code
- Start with High-Level Design: Let AI help design architecture first
- Implement in Modules: Break large tasks into manageable units
- Maintain Conversation Continuity: Keep context within the same session
- Version Control: Use Git to record AI-assisted iteration process
- Learn and Grow: Don't just use code, understand it
- Cursor - AI-native code editor based on VS Code
- GitHub Copilot - AI pair programmer for VS Code, JetBrains, Vim
- Continue - Open-source autopilot for VS Code
- Codeium - Free AI code completion
- Tabnine - AI code completion for multiple IDEs
📝 Complete AI Tools List - 50+ AI programming tools with detailed comparisons
- Claude Code - Official CLI tool from Anthropic
- Aider - AI pair programmer in terminal with Git integration
- Cline - AI assistant for terminal
- GitHub Copilot CLI - GitHub Copilot in terminal
- Bolt.new - AI-powered web development in browser
- v0.dev - AI UI component generator for React/Next.js
- Replit - AI-powered online IDE
- Codeium - AI coding assistant in browser
- LangChain - Framework for building LLM applications
- AutoGPT - Autonomous AI agent
- CrewAI - Framework for orchestrating role-playing AI agents
- OpenHands - Autonomous coding agent
- Aider - AI pair programmer with agent capabilities
📝 Agent Frameworks Guide - Complete guide to AI agent frameworks
📝 Tutorials and Learning Paths - Comprehensive tutorial collection
- Claude Code Guide - Official Anthropic documentation
- GitHub Copilot Quickstart - 15-minute getting started guide
- Cursor Introduction - Official Cursor documentation
- Prompt Engineering Guide - Anthropic's guide to effective prompting
- OpenAI Prompt Engineering - Prompting techniques for GPT-4
- Effective AI Pair Programming - Best practices for AI collaboration
📝 Video Tutorials Collection - Curated video resources
- Neural Networks: Zero to Hero - Andrej Karpathy
- AI for Everyone - Andrew Ng
- Cursor Features Deep Dive - Cursor Official
- AI Programming for Beginners - FreeCodeCamp
- Build AI Apps with OpenAI API - Traversy Media
📝 GitHub Repositories Guide - Curated list of quality repositories
- YuyaoGe/Awesome-Vibe-Coding ⭐ - Most comprehensive Vibe Coding ecosystem overview
- filipecalegario/awesome-vibe-coding - Curated list of Vibe Coding resources
- analyticalrohit/awesome-vibe-coding-guide - Become 10x Vibe Coder guide
- cpjet64/vibecoding - Collaboration space for Vibe Coders
- humanstack/vibe-coding-template - Complete full-stack template for vibe coding with code context
- anthropics/anthropic-sdk-python - Anthropic Python SDK
- anthropics/anthropic-sdk-typescript - Anthropic TypeScript SDK
- openai/openai-python - OpenAI Python library
📝 Expert Resources and Documentation - Resources from experts and major companies
- Andrej Karpathy - Former Tesla AI Director, Vibe Coding concept creator
- Andrew Ng - DeepLearning.AI founder, AI education pioneer
- Jeremy Howard - Fast.ai founder, practical deep learning advocate
- Microsoft - GitHub Copilot, Azure AI
- Google - Gemini, TensorFlow
- OpenAI - GPT-4, Codex
- Anthropic - Claude, Claude Code
📝 Open Source Models Guide - Complete guide to open source code models
- DeepSeek Coder - High-performance code generation model
- CodeLlama - Meta's code generation model
- Codestral - Mistral's code generation model
- StarCoder - Open source code generation model
- Ollama - Run LLMs locally
- vLLM - Fast LLM serving and inference
- LM Studio - Desktop app for running LLMs locally
📝 Academic research on Vibe Coding and AI-assisted programming
- Good Vibrations? A Qualitative Study of Co-Creation, Communication, Flow, and Trust in Vibe Coding - Qualitative study on collaboration, communication, flow, and trust in vibe coding
- Is Vibe Coding Safe? Benchmarking Vulnerability of Agent-Generated Code in Real-World Tasks - Evaluation of security vulnerabilities in AI-generated code for real-world tasks
- Awesome AI
- Awesome Machine Learning
- Awesome Deep Learning
- Awesome Prompt Engineering
- Awesome ChatGPT
Contributions are welcome! Please read our Contributing Guide first.
We welcome:
- New tutorials, documentation links (with brief descriptions)
- Quality video resources
- GitHub repositories (with reasons for recommendation)
- Practical experiences and case studies
- Translations and localization content
When submitting a PR, please ensure:
- Resources are accurate and verified
- Provide clear descriptions and categorization
- Follow existing format
To the extent possible under law, the contributors have waived all copyright and related or neighboring rights to this work.
Note: Vibe Coding is a rapidly evolving field. This repository is continuously updated. We recommend checking regularly for the latest updates.
Contact: Submit questions and suggestions via GitHub Issues