How-to Guides
Solve one concrete problem at a time.
If you are still learning the basics, start with Quick Start and then the Tutorial. These guides assume you already know how to submit actions and read snapshots.
Build And Debug
| Guide | Read It When |
|---|---|
| Bundler Setup | You need to configure Vite, Next.js, Webpack, or another bundler for .mel files |
| Code Generation | You want generated app-facing TypeScript from a MEL domain |
| Debugging | A submit does not do what you expected |
| Effect Handlers | You need to connect Manifesto to an API, database, or other IO |
| Re-entry Safety | An action or effect runs more than once |
Tooling Around An App
| Guide | Read It When |
|---|---|
| Developer Tooling | You want CLI checks, MEL editor support, or AI skill setup |
| Runtime Tooling Surface | You are building Studio, agent, or adapter tooling on public runtime contracts |
Add The Advanced Runtime Later
| Guide | Read It When |
|---|---|
| When You Need Approval or History | The project now needs review, durable history, or auditability |
| Advanced Runtime Assembly | You already know you need approval/history extensions and want the compact setup path |
Maintainers And Operators
| Guide | Read It When |
|---|---|
| Release Hardening | You need the current release gate, known limitations, or operator checks |
| Upgrade to Next Major | You are moving app or runtime code onto the hard-cut next-major surface |
What These Guides Assume
- You use
createManifesto()andactivate()for base runtime apps. - You submit typed actions through
app.action.<name>.submit(...). - You observe outcomes through
snapshot(),observe.state(), and runtime events. - You treat
snapshot()as the default app read model.
The advanced runtime guides add one more assumption: add approval/history packages only after the app needs review, durable history, or audit queries.
See Also
- Tutorial for the step-by-step learning path
- Integration for React and AI-agent patterns
- API Reference for package-level CLI, LSP, Studio, and runtime docs
- Architecture for the system-level model