Learn the Powerhouse stack
Powerhouse is a framework for building applications whose state evolves through operations rather than direct mutations. Instead of storing the current state, it records the ordered operations that produced it — giving every document a complete, auditable history. If you've worked with Git, event sourcing, or CRDTs, much of this will feel familiar.
This course explains how the stack is structured and how its pieces fit together. It's explanation, not step-by-step instructions — the goal is to understand what the agent builds for you and why. To ship with the agent, start at Get started; to build by hand, see Build.
Why use Powerhouse?
Because state is an operation log, teams get a set of hard problems solved by default:
- Track every change — who did what, when, addressable by index and hash.
- Sync across peers — offline edits merge deterministically, no last-write-wins.
- Build collaborative apps — concurrent editing without conflict resolution code.
- Keep auditable histories — the log is the audit trail.
- Separate business logic from UI — reducers hold the rules; editors just dispatch.
That makes it well suited to operational systems, governance tools, financial workflows, collaborative editors, and other distributed applications.
What you'll learn
By the end of this course you'll understand:
- The Powerhouse mental model — documents, state, and operations
- Actions and reducers — how state changes are described and applied
- Editors and host applications
- Connect and distributed synchronization
- Switchboard, subgraphs, and read models
- How a complete Powerhouse application fits together
The path
The chapters are designed to be read in order — each builds on the last:
Foundations → Architecture → Document models → Editors → Connect → Switchboard → Sync → Going further
Each chapter is a handful of short lessons that end with a quiz to check your understanding.
The mental model — documents, actions, and operations, and why Powerhouse is built around them.
How the Reactor, Connect, and Switchboard fit together — and where your code lives.
The typed core: documents, drives, state scopes, actions, and reducers.
React UIs that dispatch actions and render state — hooks, structure, and patterns.
The host application — how it loads document models, drives, and editors at runtime.
The server side — subgraphs that expose APIs and processors that build read models.
Replay, merge, and why conflict-free just works once your state lives in a log.
Reducer patterns under the hood, plus identity and signing on the operation log.
After the course you'll be able to navigate the Powerhouse architecture, reason about how a change flows through the stack, and read what the agent generates with confidence.
New to Powerhouse? Begin with What is Powerhouse? — every later lesson builds on the operation-log model it establishes.