Organize documentation structure - Documentation.AI

Organize documentation structure

Understand the optional Dimensions → Views → Content model and how to plan a maintainable documentation structure in Documentation.AI.

Why organization matters

Good content is not enough if people cannot find it.

In Documentation.AI, navigation is a first-class concern for both human readers and AI agents. A clear organizing model helps you:

This page introduces the core organizing model and points you to the detailed guides for dimensions, views, groups, pages, and API-specific structure.

If you work primarily from the code editor, also review Code Editor to see how the same concepts map to documentation.json.

The core model: Dimensions → Views → Content

Documentation.AI uses a three-layer navigation architecture:

Dimensions

Optional context

Views

Layout & navigation

Content

Pages, groups, API

At a high level:

  1. Dimensions (optional) define where the reader is (for example, product, version, language).
  2. Views define how navigation is arranged in that context (tabs and dropdowns).
  3. Content defines what the reader can read (pages, groups, API references).

You can:

Dimensions: define context (optional outer layer)

Dimensions are top-level context selectors that split docs into parallel variants and define where a reader is.

Use dimensions when you need parallel sets of docs that share structure but differ in content. For example:

Key rules:

Dimensions → Views → Content or Dimensions → Content.

See Dimensions for examples and design patterns.

Views: define navigation and layout

Views control how content is organized and navigated within a dimension context, or at the root when you do not use dimensions.

Views answer:

Supported view types:

Content: groups, pages, and API reference

The content layer is what readers actually consume:

See Pages for how paths work and how to move or rename pages safely.

See Groups for grouping strategies and rules.

See Organize API Reference and OpenAPI / JSON Schema Import.

You shape content structure using parent/child relationships (groups that contain pages, nested groups, or API sections) while views determine when and where that structure appears.

Content pages on containers

Any container node, groups, tabs, dropdowns, products, versions, or languages, can optionally have its own content page. This turns a purely structural node into a navigable page with its own URL.

Use content pages to create landing pages, section overviews, or hub pages that use Collection components to display child pages automatically. When a container has a content page, its label becomes a clickable link in breadcrumbs and navigation.

Key structure rules to remember

Dimensions and views

If you use dimensions, the order is always: Dimensions → Views (optional) → Content.

You cannot add dimensions inside a tab or dropdown; a view always lives inside a dimension context or at the root.

The one-child rule for containers (views)

Views obey a strict one-child rule:

You cannot mix different child types at the same level inside a single tabs or dropdowns container.

Groups and pages

A practical planning workflow

When you plan or refactor a docs set, work from the outside in.

Decide if you need dimensions

Decide which axes of variation are real, not just labels.

Design views for navigation

Within each dimension context (or at the root), decide how people move around.

Shape groups, pages, and API sections

Once views are defined, structure the actual docs.

Validate with real journeys

Before publishing, test with real workflows.

This outside-in approach keeps your organizing decisions deliberate and prevents ad hoc sidebars over time.

Example structures for common docs types

1. Simple product docs with tabs (no dimensions)

Use this when you have a single product and language, and you want a clear split between narrative guides and API reference.

Structure (conceptual):

Root (no dimensions)
└─ Views: Tabs
   ├─ Tab: Guides
   │  └─ Groups
   │     ├─ Getting Started
   │     │  ├─ Introduction
   │     │  └─ Quickstart
   │     └─ How-to Guides
   └─ Tab: API
      └─ Groups
         ├─ Authentication
         └─ API Reference (OpenAPI-backed)

2. Versioned docs using a version dimension

Use this when you maintain multiple live versions (for example, v1, v2, Beta) that differ meaningfully in content.

Structure (conceptual):

Dimension: Version
├─ Value: v1
│  └─ Views (optional) or Content
│     └─ Groups
│        ├─ Getting Started
│        └─ API Reference (v1)
└─ Value: v2
   └─ Views (optional) or Content
      └─ Groups
         ├─ Getting Started
         └─ API Reference (v2)

3. API-heavy docs with dropdowns and grouped reference

Use this when your primary surface is the API, possibly with multiple products, deployment models, or platforms.

Structure (conceptual):

Root (or inside a Dimension)
└─ Views: Dropdown (for mode)
   ├─ Option: Cloud
   │  └─ Tabs
   │     ├─ Guides
   │     │  └─ Groups
   │     │     ├─ Getting Started
   │     │     └─ Integration Guides
   │     └─ API
   │        └─ Groups
   │           ├─ Authentication
   │           └─ OpenAPI-backed Reference
   └─ Option: Self-hosted
      └─ Tabs (mirrors Cloud, with self-hosted specifics)

Best practices

What usually works well

Decide whether you need dimensions (product, version, language) before tuning view and group structure.

Once you add a tab, dropdown, or dimension value, keep it stable so links and AI references stay valid.

Aim for 2–3 levels deep: dimension (optional) → view → group → pages or API sections.

Mirror the same group and page structure across versions or languages whenever possible.

Use concise, descriptive titles that work in search results and sidebars.