Ship Before You Scale

Building a Production SaaS on a Bootstrapper's Budget.

This book targets developers who can build software but have never shipped a SaaS product to paying customers. They know React and Python. They have never touched Stripe webhooks, never configured a production deployment pipeline, and never debugged a billing edge case at midnight. This book takes them from blank repository to first paying customer, making every infrastructure and code decision visible with honest reasoning about cost, speed, and maintainability.

The through-line product is Marketflow: a farmers market and artisan market management platform. Market organizers manage vendor applications, stall assignments, recurring bookings, payment collection, and market day logistics. Free tier for markets under 20 vendors. Paid tier for larger or multi-market operations with advanced features. Every chapter builds a real feature of this product. By the end of the book the reader has a working, deployed, paying-customer-capable SaaS. The chapters are not isolated tutorials. They are a build log.

The stack: Python 3.12, FastAPI, React 18, TypeScript, PostgreSQL, Redis, Stripe, Supabase, Cloudflare (DNS, Tunnel, R2), Hetzner VPS, Coolify, Docker, GitHub Actions, Resend, Sentry, Grafana Cloud.

Five opinions run through every chapter:

Ship a working product before optimizing anything. The correct order is: working, secure, fast, scalable. Most developers invert this and build infrastructure for problems they do not have yet. A SaaS that handles 50 paying customers on a $20 Hetzner VPS is more valuable than a perfectly architected system that never launched.

Use what exists. A library that solves 90% of the problem and costs nothing is the correct choice over a bespoke implementation every time at this stage. Supabase Auth instead of rolling JWT handling. Stripe Billing instead of custom subscription logic. Resend instead of configuring an SMTP server. The remaining 10% is where you write code.

Coding assistants are a force multiplier with a failure mode. GitHub Copilot and Claude accelerate implementation significantly. They also generate confident, plausible code that skips error handling, ignores edge cases, and produces security vulnerabilities without warning. The discipline to review AI output as critically as any other code is the difference between a maintainable codebase and a liability.

The homelab is for development and staging, not production. Cloudflare Tunnel makes a homelab server accessible without a static IP or open ports. Production runs on a Hetzner VPS: predictable hardware, European data residency, and a price point that matches a bootstrapper's budget.

Stripe is the payment infrastructure. Not PayPal. Not a custom implementation. Stripe Billing for subscriptions, Stripe webhooks for every billing event. The complexity of payments is Stripe's problem. Your job is to handle the webhooks correctly.

Code examples use Python 3.12, FastAPI, React 18 with TypeScript, and the full stack listed above. Every chapter follows the same structure: the feature, the decision, the implementation, the trap, and the cost.

This book was generated using AI assistance.