# NetHackWiki

> The reference wiki for NetHack: monsters, items, dungeon levels, mechanics,
> strategy, and the major variants. Community-edited. Text is CC BY-SA 3.0.

If you want the wiki, download it. The whole thing is published in bulk and
refreshed every Friday. One request beats crawling ~17,900 pages, and the
dumps contain more than crawling can reach.

## Bulk downloads

- [Current pages](https://archive.alt.org/nethackwiki/nethackwiki_current.xml.gz): MediaWiki XML, every page's current revision. gzip, ~42 MiB.
- [Full history](https://archive.alt.org/nethackwiki/nethackwiki_full.xml.xz): every revision of every page. xz, ~48 MiB.
- [Images](https://archive.alt.org/nethackwiki/nethackwiki_current_images.tar.gz): every uploaded file. tar.gz, ~245 MiB.
- [NetHackWiki:Download](https://nethackwiki.com/wiki/NetHackWiki:Download): the above, plus ZIM/Kiwix and Mega mirrors, reader software, and how to run a local mirror.

## Crawling

- [robots.txt](https://nethackwiki.com/robots.txt): authoritative. Where it and this file disagree, robots.txt wins.
- [Sitemap index](https://nethackwiki.com/sitemap/sitemap-index-wikihackdb.xml): every URL meant to be fetched one page at a time.

Open to any ordinary client, with no account, API key, cookie or JavaScript:
every `/wiki/<Page>` article, `Category:` pages, and `Special:AllPages`. A
plain `curl` works. Send a User-Agent that says who you are.

Refused at the edge with a 403: page history, `?diff=`, `?oldid=`,
`?action=edit`, `Special:Search`, `Special:Export`, and `Special:` generally.

That gate is about capacity, not secrecy. Those are the expensive endpoints,
and they hold nothing the downloads above do not — full revision history is in
`nethackwiki_full.xml.xz`, and the wikitext of any page is in either XML file.
Take the dump instead.

## Licence

Text is [CC BY-SA 3.0 Unported](https://nethackwiki.com/wiki/NetHackWiki:Copyrights).
Image licensing varies per file; see the file's own page. Attribute
NetHackWiki and link back.