Fandom Scraper

Scrape any Fandom / Wikia wiki — Harry Potter, Marvel, Minecraft, One Piece, Genshin Impact, Star Wars, Pokémon, Naruto, and the hundreds of thousands of other communities on fandom.com — into clean, structured JSON. Built directly on the public MediaWiki Action API, so it is HTTP-only, login-free, fast, and reliable. No cookies, no browser, no breakage.

The headline feature: the article infobox (the structured side-panel of stats — a character's affiliations, an item's manufacturer, a location's region) is parsed into a flat { label: value } object. That is the part fans, datasets, and AI agents actually want, and most generic scrapers throw it away.

What you get

One JSON record per article:

Field	Description
title, url, pageId	Article identity
wiki, subdomain, language	Which wiki, and its language code
summary	Lead paragraphs as clean plain text
infobox	Portable-infobox data as a flat { label: value } object
infoboxImage	Main infobox image URL
categories	Category names the article belongs to
sections	Heading outline (level, title, anchor)
images	Image filenames used on the page
internalLinks, externalLinks	Outgoing links (capped, configurable)
wordCount, lastEdited, lastRevisionId	Size + freshness (ISO-8601 edit time)
contentText, contentHtml, wikitext	Full article body — opt-in toggles

By default records are lean (summary + infobox + metadata) so they fit comfortably in an AI agent's context window. Flip includeFullText, includeHtml, or includeWikitext when you need the whole body.

Four modes — auto-detected from your input

1. Article URLs — paste exact pages into startUrls, e.g. https://harrypotter.fandom.com/wiki/Wand. You can mix wikis and languages in a single run.
2. Search — set wiki (harrypotter) + searchQuery (elder wand); the top matching articles are scraped, up to maxResults.
3. Category — set wiki + category (Wands); every article in the category is scraped, with automatic pagination.
4. Titles — set wiki + titles (["Wand", "Elder Wand"]) when you already know the exact pages.

wiki accepts either a subdomain (harrypotter) or a full host (harrypotter.fandom.com). Language wikis work too, via either the subdomain form (es.naruto) or a language-path URL (harrypotter.fandom.com/de/wiki/Zauberstab).

Example input

{
  "wiki": "harrypotter",
  "category": "Wands",
  "maxResults": 50
}

Example output (trimmed)

{
  "wiki": "harrypotter.fandom.com",
  "subdomain": "harrypotter",
  "language": "en",
  "title": "Wand",
  "url": "https://harrypotter.fandom.com/wiki/Wand",
  "summary": "A wand was a quasi-sentient magical instrument through which a witch or wizard channelled their magical powers…",
  "infobox": { "Manufacturer": "Ollivanders, Gregorovitch", "Usage": "Casting spells" },
  "infoboxImage": "https://static.wikia.nocookie.net/harrypotter/images/5/5c/Wands-Woods.png",
  "categories": ["Wands"],
  "sections": [{ "level": 2, "title": "History", "anchor": "History" }],
  "wordCount": 10124,
  "lastEdited": "2026-05-24T19:02:03Z"
}

Who uses this

• Dataset & ML teams building knowledge bases or fine-tuning corpora on fictional universes, games, and media.
• Content & SEO sites that aggregate game guides, character bios, item stats, and lore.
• Researchers & analysts studying fan communities, franchises, and worldbuilding.
• AI agents that need a single, canonical fact source for a game or show, returned as structured JSON.

Pricing

Pay-per-event — you pay only for articles actually delivered:

• Actor start — $0.00005 per run
• Article scraped — $0.005 per article successfully returned

A 50-article category run costs about $0.25. Failed or missing pages are never charged.

For AI agents (MCP-ready)

Exposed through the Apify MCP server as apify--fandom-scraper. The input is narrow (a wiki plus a query or a URL), the output is structured JSON, and the per-article cost is predictable — exactly the shape an agent wants. Use it to pull canonical facts about characters, items, locations, and events across any fan wiki.

This is not a general web crawler. It speaks only to Fandom / Wikia (MediaWiki) wikis. For other sites, use the matching scraper in the store.

FAQ

Does it need login or cookies? No. It reads only public wiki content through the official MediaWiki API.

Which wikis work? Any wiki on fandom.com (formerly Wikia), in any language.

How current is the data? Live — each run hits the wiki in real time. lastEdited tells you when the article itself last changed.

What if a page doesn't exist? It is skipped (and not charged). If a run's pages are all invalid, the run fails cleanly with a clear message.

Can I get the full article text? Yes — set includeFullText for clean plain text, includeHtml for rendered HTML, or includeWikitext for the raw MediaWiki source.

How do I keep records small for an LLM? Leave the content toggles off (the default) and lower maxLinks.

Notes & legal

• Reads only public wiki pages via the official MediaWiki Action API; redirects are followed automatically.
• Fandom content is community-authored and licensed under CC-BY-SA — attribute the source wiki when you republish.
• Use responsibly and within Fandom's Terms of Use. This tool retrieves publicly available pages only and performs no login or circumvention.