Skip to content

README.md

Latest commit

 

History

History
66 lines (48 loc) · 3.72 KB

README.md

File metadata and controls

66 lines (48 loc) · 3.72 KB
Raw

Vibe Coding Tutorial: Building Serverless OpenClaw

How one developer built a full-stack serverless AI agent platform on AWS — for $1/month — using conversational AI coding with Claude Code.

What Is This?

This tutorial reconstructs the real development journey of Serverless OpenClaw, built entirely through "vibe coding" — conversational prompts to an AI coding assistant (Claude Code). Every prompt, decision, failure, and fix is documented from actual conversation logs.

You'll learn:

  • How to direct an AI assistant to build production infrastructure
  • AWS serverless architecture patterns (CDK, Lambda, Fargate, DynamoDB)
  • Real debugging stories — what breaks and how to fix it
  • Cost optimization techniques that reduced hosting from ~$18/month to ~$0.23/month

Prerequisites

  • Basic TypeScript/Node.js knowledge
  • AWS account with CDK bootstrapped
  • Claude Code CLI installed (npm install -g @anthropic-ai/claude-code)
  • Familiarity with AWS services (Lambda, DynamoDB, API Gateway) is helpful but not required

Table of Contents

# Chapter Time Key Topics
1 The $1/Month Challenge ~2 hours PRD, architecture design, cost analysis
2 MVP in a Weekend ~8 hours 10-step Phase 1, CDK stacks, TDD
3 Deployment Reality Check ~4 hours Docker, secrets, auth, first real deploy
4 The Cold Start Battle ~6 hours Docker optimization, CPU tuning, pre-warming
5 Lambda Migration ~4 hours Phase 2, embedded agent, S3 sessions
6 Smart Routing ~3 hours Hybrid execution, cold start preview
7 Release Automation ~2 hours Skills, parallel review, GitHub releases
8 Appendix: Cost Summary — Full cost breakdown and comparison

Total estimated time: ~29 hours across 5 weeks (Feb 8 – Mar 19, 2026)

Timeline

Feb 8-9   ████░░░░░░░░░░░░░░░░  Design & planning
Feb 9-11  ████████░░░░░░░░░░░░  Phase 1 MVP (10 steps)
Feb 12-13 ████████████░░░░░░░░  First deployment + debugging
Feb 13-16 ████████████████░░░░  Cold start optimization (#2-#9)
Feb 25-26 ████████████████░░░░  OpenClaw deep analysis
Mar 14-15 ██████████████████░░  Phase 2 Lambda migration + releases
Mar 18-19 ████████████████████  Smart routing + session continuity

How to Read This Tutorial

Each chapter follows a consistent structure:

  • Context — What we're trying to achieve
  • The Prompt — The actual prompt given to Claude Code (the "vibe")
  • What Happened — Step-by-step narrative of the AI's work
  • The Result — Output, metrics, what was produced
  • Lessons Learned — Insights and gotchas
  • Try It Yourself — Reproducible commands

Note: Prompts were originally in Korean. They are shown verbatim with English translations where helpful.

Generated from Claude Code conversation logs on 2026-03-29. Last log analyzed: session 787dc3e3 (2026-03-19)