The Comprehensive Vibecoding Guide

From a practitioner, for practitioners

A compendium drawn from real commercial projects and hundreds of thousands of lines of AI‑assisted code. Read it end‑to‑end or feed this repo to your AI agent for summaries and Q&A. Stars and watches appreciated! ⭐

Also, this guide is basically a collection of my thoughts - you may disagree with them :)

Main thought - as long as certain steps are followed - the LLM behind used for vibecoding itself doesn't matter as much as advertised - as with proper technique and overall setup even 'worse' LLMs than frontier models will be able to deliver whatever you need (at least in terms of web development).

Especially in the world of frontier models being expensive - usually it makes no sense to pay a lot just for a model being a tiny bit better than opensource (or even free) models

---

Start here

• Introduction ✨
• Quickstart 🚀

Contents

• Development Tools 🛠️
  • Recommended Tools: Zed • Droid CLI • OpenSpec • Warp • Claude Code CLI
  • MCP Servers: Context7 • DevTools • Sequential Thinking • Task Manager • Shadcn
  • Additional: Compatibility Guide
• Honorable Mentions 🏆
  • Free & Cost-Effective: Qwen Coder • Gemini CLI • AmpCode • TRAE • Windsurf • Octofriend
  • Native Integration: GitHub Copilot
  • Tools I Dropped: Traycer • GitHub Speckit
• AI Model Providers 🤖
  • Primary Providers: GLM Coding Plan • Factory AI • Synthetic.new
  • Honorable Mentions: Budget platforms (Chutes.ai • OpenRouter • nanoGPT) • Over-expensive options (Claude Subscription)
• Core Technologies 🧰
  • Astro • Tailwind CSS • Cloudflare Pages
• Hosting Tools 🌐
  • Free & Scalable Hosting: Cloudflare Pages • Workers • R2 Storage • D1 Database • KV
  • Complete edge platform for building and deploying production apps
• Business Model 💼
  • Real Income Strategy: Building websites for local businesses
  • Reality Check • The Model • Pricing & Economics • Value Proposition
• Context Management 🧠
• Workflow & Process 🔄
  • Phase 0 (Vibecoder Preparation) + Phase 1–4 deep dives
• Mastering AI Prompts 🎯
  • Foundations: Anatomy of good prompts • Universal principles • Anti-patterns
  • Task-Specific: Feature dev • Debugging • Refactoring • Code review • Testing
  • Advanced: Multi-step workflows • Prompt chaining • Error recovery • Optimization
  • Templates: 17 ready-to-use prompt templates for common scenarios
• Troubleshooting Guide 🔧
  • Quick Reference: 40+ symptom → solution lookup tables
  • Emergency Flowcharts: AI producing garbage • Can't debug • High costs • Breaking changes
  • Common Issues: AI behavior • Code quality • Workflow bottlenecks • Tool problems
  • Integrated: Phase-specific troubleshooting in workflow docs
• Glossary 📚
• Contributing 🤝

---

🧭 Learning Path

New to Vibecoding:

1. Introduction → Understanding the approach
2. Quickstart → Getting started with Git and basic setup
3. Phase 0 → Tool selection and mindset

Building Your First Project:

4. Phase 1: Planning → Project structure and specifications
5. Phase 2: Development → Implementation workflow
6. Core Technologies → Recommended tech stack

Scaling and Optimization:

7. Development Tools → Advanced tool configuration
8. AI Model Providers → Optimizing AI assistance
9. Phase 3-4 → Testing, debugging, and deployment

Improving Your Skills:

10. Mastering AI Prompts → Effective communication with AI
11. Troubleshooting Guide → Solving common problems

Business Focus:

12. Business Model → Making real money with vibecoding
13. Hosting Tools → Cost-effective infrastructure
14. Context Management → Efficient workflows

Quick Reference:

• Glossary → Key terms and definitions
• Troubleshooting → Quick problem→solution lookup
• Prompt Templates → Copy-paste ready prompts
• Contributing → How to contribute to this guide

---

Tip: Keep context lean. Link only relevant files in your prompts and use MCPs (e.g., DevTools, Context7) for efficient debugging and documentation lookup.

---

🫶 Support the Author

If you've found this guide helpful and want to support my work, consider using these referral links for the tools I recommend:

• Synthetic.new - Generous, 5h limit, privacy-first AI provider with extensive open-sourcemodel library and competitive pricing ($10 first month on standard plan)
• GLM Coding Plan - Cheap and reliable provider of a coding plan with generous 5h limit and GLM model(s) support
• Droid CLI / Factory AI - The ultimate coding agent with intelligent planning and safe execution

Using these links helps me continue maintaining and expanding this guide while you get access to excellent tools. No pressure though - the guide will always remain free and open source! ⭐

Star History