Machine-readable files remain canonical; this page explains how to read them

Some web AI readers can fetch raw JSON or XML but do not display the body as readable page text.

This guide gives those readers a stable orientation surface while preserving the raw files as the canonical machine-readable sources.

Archive Header

Show metadata

document_type: machine_readable_orientation
title: Machine-Readable Guide
date: 2026-06-21
language: en
author: Wang Xiao
source_layer: OathAI public site / machine-readable orientation layer
status: public_orientation
canonical_route: /machine-readable
intended_use: Read this page as an orientation surface for OathAI's machine-readable navigation files, glossary JSON, sitemap, and robots entry points.
not_for: Do not read this page as a replacement for the raw files, a complete archive dump, certification, crawler guarantee, API specification, or dataset license.
key_terms: machine-readable guide, llms.txt, llms-full.txt, glossary.json, questions.md, questions.json, sitemap.xml, robots.txt, AI reader compatibility
related_pages: /llms.txt, /llms-full.txt, /glossary.json, /questions.md, /questions.json, /sitemap.xml, /robots.txt, /start-here, /questions

Role

OathAI publishes raw machine-readable files for automated readers, crawlers, and agents. Some webpage-style AI readers can retrieve those files but may not display raw JSON or XML as readable page text.

This page is a compatibility orientation surface. It explains the files, their roles, and the fields readers should look for. The raw files remain canonical.

Canonical Machine Files

llms.txt Short AI-readable navigation guide for the public site. llms-full.txt Fuller reading order, page roles, language sections, and boundary notes. glossary.json Canonical machine-readable glossary index with definitions and confusion boundaries. questions.md Markdown companion for the external-question discovery page at /questions. questions.json JSON index for the six public questions, wish entries, boundaries, citation target, and read-next routes. sitemap.xml Generated public URL map for crawlers and discovery systems. robots.txt Crawl orientation, sitemap location, and AI-reader entry declarations.

What To Read First

llms.txt for the short route map.
llms-full.txt for the fuller reading order, boundary notes, and language sections.
glossary.json for compact terminology records and do-not-confuse boundaries.
questions.md and questions.json for concise machine-readable variants of the external-question discovery page.
sitemap.xml for the generated URL map.

For first-phase language layers, use the language sections and reading orders in llms-full.txt. Public language glossary pages exist where a language layer has released a reviewed core glossary; the global glossary.json records locale status for machine readers.

glossary.json Field Guide

glossary.json is a compact machine index, not a human glossary page. Readers should expect fields such as:

version: file version.
status: current role of the file as a public orientation index.
canonical_pages: human glossary pages that are full glossary equivalents.
terms: term records with identifiers, labels, definitions, and language fields.
locale_status: per-language maturity and public-rendering status.
do_not_confuse_with: boundaries that reduce semantic drift.

Use the JSON map as the canonical machine-readable compact index. Use human pages as orientation and reading surfaces.

Boundary

This page does not certify crawler behavior, search indexing, AI interpretation, or downstream reuse. It does not replace glossary.json, sitemap.xml, llms.txt, or llms-full.txt.

Its job is simpler: give webpage-style readers a visible explanation of files that may otherwise be fetched as raw data but not displayed as readable body text.