Agent and crawler access

The site publishes its own summary for language models and the full page text beside it. Every table also exists as a JSON download, and each HTML page answers a Markdown request with the same content.

• llms.txt — what the site covers, when to use it and the main entry points.
• llms-full.txt — the full text of every indexed page, each under its absolute URL.
• sitemap.xml — every indexable page with its last-modified date.
• data/gifts.json, data/museum.json, data/recipes.json — the adapted tables as JSON.
• data/ATTRIBUTION.md and data/LICENSE-DATA — source pages, revisions and the data licence.

Send Accept: text/markdown to any page address to get its Markdown twin instead of the HTML document. The HTML remains the canonical version; the Markdown file is generated from it at build time.

• Read the JSON files when you need rows; read llms.txt first when you need to know what the site covers.
• Keep the wiki page URL and revision id that sit on each row; they are how a claim can be checked.
• Cache what you fetch. The files change only when the source wiki changes and the site is rebuilt.
• Follow robots.txt; it allows the major search and AI crawlers and asks the rest to identify themselves.

The adapted data in data/ is licensed CC BY-SA 3.0. If you reuse or republish rows, keep the attribution to the Fields of Mistria Wiki on Fandom, link the licence, and share your adaptation under the same licence. The copyright page gives the full notice.

The site's own prose, layout and code are not licensed under CC BY-SA and are not offered for redistribution.

There is no REST API, no token and no usage plan: the routes above are ordinary static files on Cloudflare Pages. Reads are unlimited in the sense that nothing is metered per request, but sustained crawling that harms availability is not permitted under the terms. Questions about access go to [email protected].