ark-wiki-mcp

An MCP server that lets Claude answer ARK: Survival Ascended / Evolved questions — taming requirements, crafting recipes, spawn locations, base stats — by querying the ARK community wiki directly, instead of relying on training data or generic web search.

Tools

• search_wiki(query) — fuzzy title search, useful if you're unsure of exact spelling.
• get_creature_info(name) — diet, temperament, tameable/breedable/rideable, which maps it spawns on, saddle, base stats. Backed by the wiki's structured Cargo data.
• get_crafting_recipe(item) — ingredients + quantities, crafting station(s), engram level and point cost. Backed by structured Cargo data.
• list_wiki_sections(title) — see what sections exist on a wiki page.
• get_wiki_section(title, section) — fetch prose sections not covered by the structured tools: exact taming food/narcotic amounts ("Taming Food", "KO Strategy"), spawn locations/habitat, drops, breeding notes, etc.

When a page has separate data for ARK: Survival Evolved vs. Survival Ascended, the Ascended version is preferred automatically where the wiki distinguishes them.

Install as a Claude Code plugin

/plugin marketplace add chelseyklein/ark-wiki-mcp
/plugin install ark-wiki-mcp@ark-wiki-mcp

Requires Python 3.10+ available as python on your PATH. Dependencies (requests, beautifulsoup4, mcp) are installed automatically on first run.

Run standalone

git clone https://github.com/chelseyklein/ark-wiki-mcp
cd ark-wiki-mcp
python -m venv .venv
.venv/Scripts/pip install -r requirements.txt   # .venv/bin/pip on macOS/Linux
claude mcp add ark-wiki -- .venv/Scripts/python.exe server.py

How it works

Structured facts (creature stats, crafting recipes) come from the wiki's Cargo tables via its public API — no scraping. Prose content (taming strategy, spawn habitat descriptions) is fetched per-section via the MediaWiki API and converted from HTML to plain text.

License

MIT