JSON Schema for Humans vs Natural Docs

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`.

Natural Docs is a command-line documentation generator that scans source-code comments and text files to create HTML documentation, including API references. Its syntax is designed to read like ordinary English in both source code and generated pages. The project says it comes configured for 25 programming languages and can be extended; for certain languages, it can also read Javadoc and Microsoft XML comments without rewrites. Generated pages have three panels, dynamic navigation, search and summaries that appear when readers hover over links. They can be opened from a local hard drive without installed software, and styled with custom CSS and images. Natural Docs runs on Linux, macOS and Windows. It is free and open source under AGPL version 3. The license requires source availability for modified distributions, projects incorporating its source, and services incorporating its functionality; merely documenting a project does not place that project under AGPL. Help is available through a Reddit forum, email and GitHub issues, with commercial support contracts available to businesses.