Beginner — Getting Started
This guide is written for people who are not full-time developers. e2e-base has a different grain from the other ddakit bases. You are not building an app — you are standing up a test suite that checks, automatically, whether an app you already have behaves the way you described.
So when this guide ends, what you hold is not a new app. It is an automated check that runs flows like “does login work, does checkout go through, does the core feature save” in one pass, instead of a person clicking through them by hand every time.
Who this guide was written for
Section titled “Who this guide was written for”A fair share of the people who pick up this base are not full-time developers, and that shaped the tone — it walks alongside you. There is one assumption underneath it, though.
You have opened a terminal at least once and have written code alongside an AI. Commands get written out one line at a time, but this guide does not teach what a terminal is, what git is, or why Node is needed from the ground up. Keep a free primer in a side tab and lean on it as you go.
Three things are assumed.
- You recognize terminal, git, and Node by name and have a rough sense of what each is for. The guide writes commands out a line at a time on top of that.
- You already have an app to test. This base does not build apps. A website, a desktop app, or a phone app you made has to be working already. e2e-base knocks on that app from the outside.
- The test code itself gets written by an AI. Claude Code is what this guide uses. You only need to know which flows would be fatal if they broke.
If a step already feels familiar, skim it. It still pays to glance at each one once, so when something behaves differently than you expected you know where to come back.
| Step | What you do |
|---|---|
| Prerequisites | Install Node, git, a package manager, and Claude Code. Track-specific test tools come after you pick a track |
| Clone and install | Get the base and bootstrap with bash init.sh |
| First run | One line — start project setup — detects the stage and sets the pace |
| Choose your target | Decide whether your app is web, desktop, or mobile, and pin down the flows that would be fatal if they broke |
| Your first test | The seat where one critical-flow line becomes one test |
| Run and read the report | Run the tests and read the result in flow language |
| The automated safety net | The hooks and validation gates the base runs for you |
| When the AI gets stuck | Unblock the workflow itself when it loops in the same spot |
| Troubleshooting | A collection of common stuck points |
| Staying up to date | Pull base improvements into your own copy |
If Playwright or Maestro already feel familiar, the Expert manual covers the same base at a shorter pace.
The one thing you must never do
Section titled “The one thing you must never do”When you point the suite at the app under test, never point it at a live production service or a real database. Point it only at local, staging, or a dedicated test environment.
That single line is the most important safety boundary in this base. The tests attempt signups, click checkout, create and delete data. If the target is a server real users depend on, real data gets corrupted, real charges fire, and real accounts can be deleted. This seat gets reinforced again in Choose your target and The automated safety net.
When something goes wrong
Section titled “When something goes wrong”The first place to look is Troubleshooting. Almost every common first-run stuck point is collected there.
If your symptom is not on the list, copy the error message verbatim into Claude Code and ask “how do I fix this?”. A next thing to try usually comes back.
If the AI loops on the same spot twice — in the workflow itself rather than the code — that belongs to When the AI gets stuck. The two pages cover different kinds of stuck.
License
Section titled “License”The manual itself is free to read, open in a side tab as you go. The actual base source and the full .claude/ payload (agents, skills, hooks) ship under a commercial license. When you reach the point in the manual where you start wondering “where does the code with all this automation actually live?”, that is the seat where you take a license and fork.