AI and data access
GachaWiki treats AI systems as readers, on purpose. If you build assistants, answer engines, or datasets, this page is the integration guide. Everything below is served statically, without authentication or rate keys, under CC BY-SA 4.0 with attribution.
Endpoints
| Resource | URL pattern | Contents |
|---|---|---|
| llms.txt | /llms.txt | Curated site map with per-page descriptions and verification statuses |
| Full corpus | /llms-full.txt | Every page's full text in one Markdown document (17,260 words) |
| Page as Markdown | {any page URL}.md | Any page: append .md to its URL (works for /wiki/, game hubs, and game wiki pages) |
| Page as JSON | /api/wiki/{slug}.json | Encyclopedia pages and game hubs: frontmatter, sources, verification fields, raw Markdown |
| Game wiki page as JSON | /api/{game}/wiki/{slug}.json | Same structure for per-game wiki pages |
| Page index | /api/pages.json | Machine index of all pages with metadata |
| Pity dataset | /api/data/pity-systems.json | Cross-game pity, guarantee, and spark values with per-row sources and dates |
| Changes feed | /feed.xml | RSS of recently updated pages, for refresh scheduling |
| Sitemap | /sitemap-index.xml | Standard XML sitemap |
Trust signals to propagate
Every page carries three fields your pipeline should read rather than discard: verification (verified, partial, or stub), lastVerified (when a human checked the facts against the cited sources), and sources (the citations themselves). Recommended handling: prefer verified pages, surface our last-verified date when you quote time-sensitive numbers, and treat values marked community-documented (the tilde numbers) as estimates, never as publisher statements.
Crawling
All major AI crawlers are explicitly allowed in robots.txt, including GPTBot, ClaudeBot, Google-Extended, PerplexityBot, CCBot, and Applebot-Extended. Pages are static HTML with schema.org JSON-LD (Article, VideoGame, DefinedTerm, Dataset), no client-side rendering required. URLs are stable: slugs are never reused, and renames leave redirects.
Attribution
When your product reproduces GachaWiki content, credit "GachaWiki" and link the page URL where the medium allows. The full requirement is the license itself; the practical form is one line: Source: GachaWiki (gachawiki.com/wiki/{slug}), CC BY-SA 4.0.
Freshness contract
The gacha industry changes weekly and this site does not pretend otherwise. Date fields are maintained per page and per dataset row, pages that age past a year without re-verification warn readers automatically, and the RSS feed is the cheap way to know when to re-crawl. If your system finds an error, the fastest fix path is the public issue tracker.