Tarot MCP Server

@pipeworx/tarot

Draws, searches and spreads from an original 78-card tarot deck written for this pack, with seeded/reproducible randomness, optional card reversals, named spreads, and stateful shuffle sessions.

Part of Pipeworx — an MCP gateway connecting AI agents to 1866+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.

What changed

This pack no longer calls tarotapi.dev. The 78-card deck — every card's name, keywords, upright/reversed meaning and image description — is written for this pack and shipped inline (src/deck.ts, ~46 KB); nothing is fetched at call time. random_card and draw_cards keep their original argument shapes and gain seed, reversals, and session_id. Two tools are new: tarot_spread (named layouts) and tarot_shuffle (sessions).

Tools

• random_card(seed?, reversals?, session_id?) — draws one card. Returns the card, its orientation (upright/reversed — a 50% coin flip unless reversals: false), the traditional meaning for that orientation, and the seed used (pass it back to reproduce the exact draw).
• draw_cards(count, seed?, reversals?, session_id?) — draws 1–78 cards the same way. (The gateway exposes this tool as tarot_draw_cards when it collides with another pack's draw_cards; the pack-local name stays draw_cards.)
• search_cards(query) — searches the deck by card name, suit, or keyword (either orientation). A query that matches nothing is a loud error, not an empty list.
• get_card(name_short) — looks up one card by id (major_00..major_21, wands_01..pentacles_14) or exact name ("The Fool", "Ace of Wands"). Returns both the upright and reversed meaning as reference.
• tarot_spread(spread, positions?, question?, reversals?, seed?, session_id?) — deals a named layout: three_card, situation_obstacle_advice, celtic_cross (10 positions), horseshoe (7 positions), yes_no (1 position), or custom (caller supplies positions[], max 20). Each position gets its own card, orientation and meaning. question is echoed back verbatim — this tool never interprets it; the calling agent does.
• tarot_shuffle(session_id?, seed?) — opens a new 78-card session, or reshuffles an existing one if session_id is given.

Meanings are traditional; interpretation is the calling agent's job, not this pack's — every tool description says so.

Sessions are hosted-only

session_id (on the drawing tools and tarot_spread) and tarot_shuffle need the gateway's shared KV to track which cards have already been drawn. A standalone/npm build of this pack has no KV binding, so those calls return { session_supported: false, hint } instead of throwing — single-call seeded draws work identically either way, since they need no session.

A session draws without replacement across calls (it never repeats a card until reshuffled), expires 24h after it's opened or last reshuffled, and its card order is reproducible from the seed tarot_shuffle returns.

Auth

Keyless. No upstream, no API key, no rate-limited third party — the deck is inline and the randomness is local.

Data sources

This pack has no live data source. The deck's structure (suit/rank layout, the Major Arcana's traditional planet/zodiac correspondences, the suits' elemental correspondences) follows centuries-old public-domain tarot tradition. The Rider–Waite–Smith deck and A.E. Waite's 1910 "Pictorial Key to the Tarot" (public domain) were consulted only as a reference for which traditional images and themes this deck tracks — every card's keywords, meanings and image description in src/deck.ts are original text written for this pack, not a transcription of Waite's or any other published deck's booklet.

Reproducing a draw

Every drawing tool returns meta.seed. Passing that seed back as the seed argument on an otherwise-identical call reproduces the exact same cards and orientations. Omit seed for fresh randomness — the seed used is always returned either way, so a draw is always citable even when it wasn't requested as reproducible.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "tarot": {
      "url": "https://gateway.pipeworx.io/tarot/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/tarot/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1866+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

curl -X POST https://gateway.pipeworx.io/v1/tools/tarot_random_card \
  -H 'Content-Type: application/json' \
  -d '{}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/tarot_random_card. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.

Run it from npm

@pipeworx/mcp-tarot is a local stdio bridge to the Pipeworx gateway: your MCP client starts it with npx, and it forwards every call to the same https://gateway.pipeworx.io/tarot/mcp endpoint shown above. It needs network access; there is no offline mode.

{
  "mcpServers": {
    "tarot": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-tarot"]
    }
  }
}

Without a key it calls anonymously, capped at 50 calls a day. A free account raises that to 200 a day: get a key at https://pipeworx.io/signup and add "env": { "PIPEWORX_API_KEY": "<your key>" } to the server entry. The key is sent as a Bearer token, the same way a hosted connection sends it, so calls are metered and billed exactly as they are there.

It lists only this pack's tools by default. Set PIPEWORX_META_TOOLS=1 to also list ask_pipeworx and the other gateway-wide tools.

Behind an HTTPS proxy: Node's built-in fetch does not read HTTPS_PROXY by default. Newer Node releases honour it when NODE_USE_ENV_PROXY=1 is set; check your Node version's documentation.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Tarot data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT