Core Concepts
Core concepts and architecture that underpin Documentation.AI.
Documentation.AI focuses on four concepts. Use these to choose workflows and architecture.
AI-native architecture
-
AI assists across the lifecycle: draft, revise, evaluate, and search.
-
Private-by-default controls and clear data handling. See AI assistant setup.
-
Tune retrieval and answers with search configuration.
-
Keep humans in control: AI suggests; you review and publish.
MDX + component system
-
Content is MDX with a rich component library.
-
Keep prose in markdown; use components for structure, interactivity, and API reference.
-
Common building blocks: callouts, code blocks/groups, tables, API request/response.
-
Prefer small, reusable sections over long monolithic pages.
Web and git workflows
-
Web editor: collaborative, visual, component-aware. See Web editor.
-
Git workflow: file-first, PR reviews, branches. See Code editor.
-
When to use: web for quick edits and non-developers; git for reviews, branching, CI.
-
Hybrid is common: authors in web; reviewers in git.
Deployment options
-
Managed hosting with global CDN, SSL, and custom domains. See Deploy your documentation.
-
Self-hosted export for custom infra and CI/CD. See Build information.
-
Public or private spaces; map subpaths or separate domains.