MVI

MVI stands for Model-View-Intent, a user interface architectural pattern that emphasizes a unidirectional data flow. It structures UI logic around a single source of truth and a predictable loop from user interaction to rendered state.

In MVI, the View renders a Model, which is an immutable representation of the current UI state. The View also emits Intents, which represent user actions, events, or lifecycle cues. These Intents are fed into a central loop that applies them to the current Model to produce a new Model. The View then renders the updated Model. The Model serves as the single source of truth for the UI, and rendering is driven entirely by Model updates.

Intents can also be referred to as Actions or Events in some implementations. The transformation from Intent to a new Model is often handled by a component such as a Processor, Reducer, or Presenter, commonly using reactive streams or asynchronous pipelines. MVI is related to but distinct from patterns like Model-View-Controller (MVC), Model-View-Presenter (MVP), and Model-View-ViewModel (MVVM), sharing with Elm and Redux the emphasis on unidirectional data flow and explicit state transitions.

Benefits of MVI include improved testability, easier reasoning about UI state, deterministic rendering, and straightforward time-travel debugging. Criticisms include potential boilerplate, a steeper learning curve, and complexity that may not be warranted for simpler applications.

MVI has been used in mobile and web development, particularly in environments that support reactive programming with languages and frameworks such as Kotlin, JavaScript, RxJava, Coroutines, or Flow. It is chosen when clear state management and predictable UI behavior are priorities.