Guides
The long-form pieces, in the order they are useful: decide what you are building on, set the agent up so it works in your codebase, then know where the approach stops working.
Start here
Start a SaaS With AI Agents in 2026
A working playbook for starting a SaaS with coding agents in 2026: what the agent should own, what you should own, and why the foundation you hand it decides the outcome.
11 min read
Choose Your Stack Before You Prompt
Letting a coding agent pick your stack mid-build is how projects end up with three ORMs and two auth systems. Decide the stack first, then prompt. Here is the decision order that works.
11 min read
Idea to Deploy in One Week
A day-by-day plan for going from idea to a deployed, billable SaaS in seven days using a boilerplate and a coding agent. What to build, what to skip, and what to cut when you fall behind.
12 min read
Decide your stack
SaaS Auth Decision Guide
Clerk, Auth.js, Supabase Auth, WorkOS, or roll your own: a decision guide for SaaS authentication in 2026, including what changes when a coding agent writes the code around it.
12 min read
SaaS Payments Decision Guide
Stripe, Lemon Squeezy, Polar, or Paddle: how to choose a SaaS payment processor in 2026 based on tax handling, not API quality, and what generated billing code gets wrong.
12 min read
SaaS Database Decision Guide
Postgres is the answer. The real database decisions for a SaaS in 2026 are hosting, ORM, and how you enforce tenant isolation, especially when an agent writes the queries.
12 min read
Set up your agent
Setting Up Claude Code for a New Project
The first hour with Claude Code on a new repo decides the next three months. A concrete setup: instruction file, commands, permissions, subagents, and the verification loop.
12 min read
MCP Servers Every SaaS Needs
Which MCP servers actually earn a slot in a SaaS development setup, which ones waste context, and why exposing your own product over MCP is becoming a distribution channel.
12 min read
CLAUDE.md vs AGENTS.md: Which File Does Your AI Agent Actually Read?
Claude Code reads CLAUDE.md. Most other agents read AGENTS.md. Here is which file each tool loads, how to keep one canonical source with a symlink, and what belongs inside.
11 min read
The AI-Agent-Ready Boilerplate Checklist: 10 Things Claude Code Needs From Your Starter Kit
In 2026, the best boilerplate is the one your AI coding agent can understand. Here are the 10 properties that make a starter kit agent-ready, and how to check them before you buy.
7 min read
Know the limits
When Not to Use a Boilerplate
We run a boilerplate directory and we still think there are cases where buying one is the wrong call. Here are the situations where a starter kit costs you more than it saves.
13 min read
The Hybrid Build: How Shipped SaaS Actually Combine AI App Builders and Boilerplates
The AI-builder-vs-boilerplate debate misses how successful products are actually built: validate on Lovable or Bolt, ship production on a boilerplate, extend with a coding agent. The full hybrid playbook.
11 min read