# The Generative UI Guide

> A plain, vendor-neutral reference on generative UI: what it is, how it works, how to build it, and how it compares to server-driven UI and A/B testing.

## Guides

- [What is generative UI?](https://generativeui.guide/): Generative UI assembles the interface for each visit at request time from live data and rules or a model, instead of shipping one fixed layout to everyone.
- [How to build generative UI](https://generativeui.guide/building): Generative UI has three parts: state that describes the visit, a decision layer that chooses what to render, and a renderer that draws it. Here is how they fit.
- [Generative UI vs server-driven UI and A/B testing](https://generativeui.guide/vs): Generative UI, server-driven UI, and A/B testing all move layout decisions off the fixed page. They differ in what decides the layout and how often it changes.
- [The generative UI ecosystem](https://generativeui.guide/ecosystem): The tools for generative UI sort into three layers: a transport, a declarative payload, and sandboxed apps that run inside an assistant.
- [How to measure generative UI](https://generativeui.guide/analytics): When every visitor sees a different interface, page views and funnels stop working. Measuring generative UI means recording what was composed, and why.
- [The history of generative UI](https://generativeui.guide/history): Generative UI is the newest round of a century-old argument about authoring systems instead of artifacts, running from typography and art to server-driven UI.

## Elsewhere

- [Snowplow Signals](https://signals.snowplow.io): Real-time customer context for generative UI and in-session personalization. Maintains this guide.
- [Generative UI on Signals](https://signals.snowplow.io/learning/generative-ui): The Signals Learning Center cluster on generative UI.

## For agents

- Every guide is served as markdown at its path with a .md suffix, for example /building.md.