@pipeworx/mcp-eu-cellar

@pipeworx/eu-cellar

Every document the Publications Office of the European Union has issued — Official Journal acts, treaties, consolidated texts, COM proposals and Court of Justice judgments — searchable by title and readable by CELEX number, from the CELLAR semantic repository's public SPARQL endpoint.

Part of Pipeworx — an MCP gateway connecting AI agents to 1864+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.

Tools

• cellar_search_works(query, resource_type?, since?, until?, limit?) — full-text title search across every EU resource type; returns CELEX, date, resource type and a direct EUR-Lex link.
• cellar_work(celex? | cellar_id?) — one document's catalogue entry: every English title, document date, in-force flag, resource types, EuroVoc subject concepts with English labels, authoring bodies.
• cellar_sparql(query, limit?) — your own SELECT or ASK against the repository. Writes are refused; results are capped.

Auth

Keyless. No registration, no key, no published quota.

Relationship to the eur-lex pack

Both talk to publications.europa.eu and they are not duplicates. eur-lex is a curated legal-compliance surface: a hand-kept index of ~14 named acts (GDPR, AI Act, DORA…), article-level text extraction and CJEU judgment lookup, with its SPARQL use hard-wired to legislation. This pack is the repository itself — all resource types, plus an arbitrary SELECT-only query tool.

Use eur-lex for "what does GDPR Article 17 say". Use this for "what has the EU published about X" and for graph questions nobody wrote a tool for.

Data sources

• https://publications.europa.eu/webapi/rdf/sparql — the CELLAR SPARQL endpoint (Virtuoso), serving the Common Data Model (cdm:) graph.
• http://publications.europa.eu/resource/cellar/{id} — the Work URIs the endpoint returns.

Traps, all verified live 2026-09-17

• Full text needs bif:contains, not FILTER(CONTAINS(...)). The endpoint is Virtuoso; a CONTAINS filter over ?title scans every expression in the repository and times out. The full-text form is ?title bif:contains "'phrase'" — note the INNER single quotes, which Virtuoso requires for a multi-word phrase.
• CELEX literals are typed. ?w cdm:resource_legal_id_celex "32024R1689" matches nothing; it needs ^^xsd:string. This is a clean HTTP 200 with zero rows, not an error.
• Titles come back both with and without a language tag. Filtering on lang(?title) = "en" silently drops a large share of the repository. Filter on the expression's language URI instead — that is the modelled fact.
• One Work returns many rows. The (title × type × subject × author) cross product means a single document legitimately comes back as 20 bindings. cellar_work collapses them; hand-written SPARQL will not.
• format= is a query parameter. The Accept header alone is honoured inconsistently and you can get RDF/XML where you asked for JSON.
• A malformed query is HTTP 200 with an HTML error page. Parsed naively that reads as "no results". This pack raises it as a rejected query.

Quick Start

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

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

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/eu-cellar/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 1864+ 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/cellar_search_works \
  -H 'Content-Type: application/json' \
  -d '{"query":"carbon border adjustment","limit":4}'

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

Run it from npm

@pipeworx/mcp-eu-cellar 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/eu-cellar/mcp endpoint shown above. It needs network access; there is no offline mode.

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

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 Eu Cellar data" })

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

More

• Docs and guides
• pipeworx.io

License

MIT