Write domain transition rules in MEL. Core computes semantic transitions, Host converges snapshots, and SDK exposes the app-facing runtime.
const app = createManifesto(schema).activate()
The same domain, snapshot, and action input produce the same next snapshot.
MEL declares domain rules, Core computes semantic transitions, and Host fulfills declared effects through an explicit loop.
Submit typed actions, read the updated Snapshot, and give UI, backend, and agent code the same state contract.
domain Counter {
state {
count: number = 0
}
computed doubled = count * 2
action increment() {
onceIntent {
patch count = count + 1
}
}
}| Step | Read | You Can Build Afterward |
|---|---|---|
| 1 | Quick Start | One MEL domain running through the SDK |
| 2 | Project Anatomy | Know where domain, runtime, UI, and agent files go |
| 3 | MEL Domain Basics and MEL For App Developers | Understand the .mel file before it grows |
| 4 | Creating an App | Activate the runtime and read snapshots from TypeScript |
| 5 | Tutorial | Build a small Todo app before adding a UI framework |
| 6 | Bundler Setup and Code Generation | Emit the generated Todo facade from the same .mel file |
| 7 | React and Runnable Examples | Connect a browser UI and compare it with a working app |
| 8 | Web App + Agent | Put UI and agent writes behind one shared server runtime |
| 9 | AI Agents | Go deeper on agent-only tool loops and server workers |
Stay on the base SDK runtime until the app path is clear. Add approval or history only when writes need review, actor records, or durable history.
Open those pages when a guide sends you there.