LLM & Agent Integration

The Wiki exposes plain-text endpoints. Models and agents can fetch Markdown directly instead of parsing HTML.

Text endpoints

curl 'https://wiki.anibt.net/docs/apply.md'

Markdown on the main site

Public HTML pages on AniBT (anime pages, release details) support content negotiation. Send Accept: text/markdown to get cleaned Markdown.

curl 'https://anibt.net/anime/638151' \
  -H 'Accept: text/markdown'

JSON APIs, RSS feeds, and static files keep their own formats.

Crawling and indexing the main site

Search engines and other crawlers find AniBT's public pages through robots.txt and the sitemaps. Every indexable page has three language URLs: Simplified Chinese without a prefix, Traditional Chinese under /zh-hant/…, and English under /en/…. Each language URL declares itself as canonical instead of pointing back to the Simplified URL. Page content never includes magnet links, torrent download URLs, or infohashes; download addresses are fetched only when you click a button.

Indexable pages

The table lists the Simplified URLs without a prefix; the /zh-hant/… and /en/… forms follow the same rules.

On /magnets and the group Releases tab, previous and next are plain links that carry the position in the URL's cursor parameter, so every page can be reached without running JavaScript. An expired page link returns to the first page for the same filters.

Language versions

An indexable page lists its three language URLs in the HTML <head> (hreflang zh-Hans, zh-Hant, and en), itself included, plus x-default pointing to the Simplified URL. The three URLs share the same path and query parameters.

<link rel="canonical" href="https://anibt.net/en/anime/617123"/>
<link rel="alternate" hrefLang="zh-Hans" href="https://anibt.net/anime/617123"/>
<link rel="alternate" hrefLang="zh-Hant" href="https://anibt.net/zh-hant/anime/617123"/>
<link rel="alternate" hrefLang="en" href="https://anibt.net/en/anime/617123"/>
<link rel="alternate" hrefLang="x-default" href="https://anibt.net/anime/617123"/>

Noindex pages (searches without a keyword or without results, /magnets with a q keyword), error pages, the sign-in pages, and pages that require signing in list no language alternates.

robots.txt

https://anibt.net/robots.txt does not block /magnets or /search, nor their /zh-hant/ and /en/ forms. Under /api/, only /api/torrent/ and /api/other-torrent/ may be crawled; other API paths, /rss/, and the sign-in and admin paths are blocked. The /zh-hant/ and /en/ forms of the admin, sign-in, and sign-up paths (for example /en/admin/, /zh-hant/auth/, /en/sign-in) are blocked too; APIs and RSS have no language prefix.

Sitemaps

https://anibt.net/sitemap.xml is a sitemap index that lists the URL sets below directly, without nesting another index.

On release URLs, lastmod is when the release's content last changed; it is omitted when that time is unknown. A month the index does not list returns 404. The old /sitemap-anime.xml still works, so a webmaster tool that already has it needs no change, but the root index no longer refers to it.

Every URL set lists each page three times: one <url> each for the URL without a prefix, the /zh-hant/ form, and the /en/ form. Each carries the same xhtml:link language alternates as the page head, and the three share one lastmod.

<url>
  <loc>https://anibt.net/release/rel_-OH17wI8Sghm</loc>
  <lastmod>2026-09-30T16:04:00.592Z</lastmod>
  <xhtml:link rel="alternate" hreflang="zh-Hans" href="https://anibt.net/release/rel_-OH17wI8Sghm"/>
  <xhtml:link rel="alternate" hreflang="zh-Hant" href="https://anibt.net/zh-hant/release/rel_-OH17wI8Sghm"/>
  <xhtml:link rel="alternate" hreflang="en" href="https://anibt.net/en/release/rel_-OH17wI8Sghm"/>
  <xhtml:link rel="alternate" hreflang="x-default" href="https://anibt.net/release/rel_-OH17wI8Sghm"/>
</url>
<!-- Followed by /zh-hant/release/rel_-OH17wI8Sghm and /en/release/rel_-OH17wI8Sghm with the same alternates -->

Deleted releases

After a release is deleted, its detail page in all three languages and its torrent download return 404:

• Anime releases: /release/<id>, /zh-hant/release/<id>, /en/release/<id>, /api/torrent/<id>.torrent
• Other releases: /other-release/<id>, /zh-hant/other-release/<id>, /en/other-release/<id>, /api/other-torrent/<id>.torrent

When the deletion receipt is completed (see Deletion status), all four URLs have been confirmed to return 404. The sitemaps then stop listing the release, and IndexNow is notified of the deletion of the three detail page URLs.

IndexNow

AniBT tells search engines that take part in IndexNow about changes to release detail pages, submitting to https://api.indexnow.org/indexnow.

• Changes are collected in 5-minute windows and each window is submitted as one batch; a URL repeated within a window is submitted once. Before submitting, AniBT checks again that the release is still public or has been deleted.
• Only release detail pages are submitted. Anime pages, group pages, and list pages are found through the sitemaps and ordinary crawling; searches, URLs with query parameters, download addresses, APIs, release previews, and pages that require signing in are never submitted.
• The key file is served at https://anibt.net/<key>.txt as text/plain.
• A 200 or 202 from IndexNow only means the submission was received, not that any page has been indexed. Google does not take part in IndexNow.