Rails Whey badge Rails Whey

Be a real maker.

Solve problems and create value with Rails and AI agents. Strong fundamentals and clear thinking help you do more and do it better.

View on GitHub92
Rails Whey tub: Strength from fundamentals

Code got cheap. Judgment didn't.

Deciding what to build, how to structure it and what to ship takes judgment. AI amplifies the foundation you bring.

Fundamentals help you plan, review and deliver. The agent writes code; you set the conditions and evaluate the result.

With fundamentals, you delegate.

You guide the agent, know what to evaluate and decide whether the solution solves the problem.

Without them, you just hope.

The code looks fine, the tests pass, and you only find the problem when a user does.

Three paths meet at Rails.

Where you start shapes your path. All three paths connect fundamentals, Rails and AI to build with judgment.

The amplifier

AI

Agents speed things up. You evaluate the result.

The structure

Rails

Conventions that organize your app and give agents context.

The foundation

Ruby, SQL, Web and infrastructure

Language, data, the Web and operations support your decisions.

Developer with a foundation

Rails or another stack. You already have the foundation.

  1. Your foundation

    Start with the experience you have.

  2. A Rails product

    Bring your experience into the app.

  3. Amplify with AI

    Delegate and review the result.

Junior developer

Start by writing code yourself.

  1. Fundamentals

    Write Ruby, SQL and Web code yourself.

  2. Build with Rails

    Put the foundation into a real product.

  3. Then, AI

    Use agents with judgment.

Vibe coder

Already build with AI? Learn the fundamentals to guide your agent and evaluate the result.

  1. Development cycle

    Plan, review and deliver with AI.

  2. Rails repertoire

    Understand what supports your app.

  3. Go deeper

    Ruby, SQL, Web and infrastructure.

    If you want to go deeper

AI amplifies the foundation you bring.

Each path combines fundamentals, Rails and AI at a different depth. The goal is to build solutions with sound judgment.

We don't claim. We measure.

We built the same app 28 ways and measured how much each version helps an AI agent do its job: context cost, discoverability, isolation, predictability and blast radius.

1A-fat-controller

10/25

agent-friendliness

Everything in one controller. To fix a 10-line bug, the agent loads 363 lines across 2 files.

6G-named-orchestrations

24/25

agent-friendliness

Every concept has a name and a home. The agent finds what it needs without getting lost.

Same app. Same tests. What changed was the judgment.

See all 28 versions

Agent-friendliness score: 5 dimensions, 1 to 5 each. Source: the Agent-Friendliness report of railswhey/app

Rails Whey App

The app behind what we teach.

A complete task app for exploring architectural decisions with Rails itself. Meet the project, see what it does, and choose where to start.

Meet the app View the code on GitHub92

Who it's for

You can start in different places and still build with judgment.

Developer with a foundation

You have
Experience with Rails or another stack.
You want
Build solutions with agents at work, for clients or in your own projects.

From experience to autonomy with agents.

Junior developer

You have
The drive to start, with a foundation still taking shape.
You want
Learn by writing code, build a real product and prepare for a first job.

From your first line to your first product.

Vibe coder

You have
An AI-built app and questions about what is behind it.
You want
Do more with AI, improve the results and learn to evaluate the solutions you create.

From an app that works to an app you understand.

You build a real product. We give you the judgment.

In the courses you build the Rails Whey app with Rails and AI agents. Every course delivers the foundation two ways.

Understand the decision

Teaches the judgment

The why behind each decision and the tradeoff that comes with it. You understand instead of memorizing it.

  • Why this decision?
  • What tradeoff comes with it?
  • How do you review the result?

Bring it into your project

Guards the judgment

Skills, CLAUDE.md, rules and checklists make the judgment explicit for the agent. You follow the work and review the result.

Instructions for the agent
# CLAUDE.md
- Use what Rails already has before hand-rolling.
- Multi-model coordination gets its own name.
- Only the 7 REST actions. If it doesn't fit, the resource is wrong.

You learn. The project keeps it. You review.

  • Recorded lessons
  • Up to 15 minutes each
  • At your own pace

Who's behind it

Serradura

Serradura

Father, Rubyist 💎, into software design & architecture. Building durable, retrievable context for humans & AI agents.

Jackson Pires

Jackson Pires

Ruby enthusiast since 2007. Loves to write software and to teach, and has trained more than 40,000 people with Ruby on Rails and programming courses.