DocFX vs JSON Schema for Humans

DocFX is a free static site generator for technical documentation. It builds documentation from .NET assemblies, XML code comments, REST API Swagger files and Markdown, and can produce rendered HTML pages, JSON models or PDFs. Static HTML output is placed in the `_site` folder for publication to hosting services such as GitHub Pages. DocFX can also be used as a NuGet library through the `Docfx.App` package and .NET APIs. It parses CommonMark-compliant Markdown with Markdig and supports additional extensions. Built-in templates include a modern option with dark mode and customization settings; users can also customize templates with CSS, JavaScript, Mustache templates and metadata. DocFX can generate git links for GitHub, GitLab, Azure Pipelines, AppVeyor, TeamCity and Jenkins. It requires .NET SDK 8.0 or higher. Node.js v20 or higher is optional generally, but required to create PDFs. Native reference documentation generation covers .NET and REST APIs; other languages need a custom API docs converter. PlantUML diagrams use a remote server by default, with local rendering available as a setting.

JSON Schema for Humans generates static HTML or Markdown documentation from valid JSON Schema files. Inputs can be JSON or YAML, supplied to its command-line interface as a file, directory, glob pattern, or comma-separated list; Python methods are also available. Multiple templates produce different outputs. The default JavaScript template provides expandable sections, tabs for conditional subschemas, and anchor links. The js_offline template bundles CSS and JavaScript for offline viewing. Markdown output can be published without an HTTP server and is rendered by GitHub when browsing code. A techdocs template is described as optimized for TechDocs and MkDocs Material. References can point within a schema, to local files, or to URLs, including circular references, and the project lists JSON Schema Draft-07 support. Safe mode escapes HTML tags in descriptions by default, while configuration can allow HTML. Property names and size, property dependencies, and media are listed as unsupported. The Apache-2.0 licensed package documents installation with `pip install json-schema-for-humans`.