Copernicus Cds MCP Server

@pipeworx/copernicus-cds

The Copernicus Climate Data Store (C3S/ECMWF) catalogue — the ~140 published climate datasets behind ERA5, CMIP6/CORDEX projections, satellite essential climate variables and seasonal forecasts — plus the ability to submit a retrieval job against any of them.

Part of Pipeworx — an MCP gateway connecting AI agents to 1683+ live data sources.

Tools

• cds_collections(query?, limit?, offset?) — list or keyword-search the catalogue. Returns collection id, title, first line of the abstract, and the temporal/spatial extent. Keyless.
• cds_collection(collection_id, include_form?) — one collection in full: STAC record (abstract, licence, keywords, extent, documentation links) plus the retrieval request form, i.e. the variables, years, levels, area and output formats the job will accept. Keyless.
• cds_submit_request(collection_id, inputs, _apiKey) — submit an asynchronous retrieval job and return the job id. Needs a caller key.

Auth

Catalogue browsing is keyless and verified so (2026-09-17). Submitting a retrieval job needs a free CDS Personal Access Token, passed as _apiKey.

No platform key is provisioned. The intended env var when one lands is PLATFORM_CDS_KEY; sign up at https://cds.climate.copernicus.eu.

platformKeyEnv is deliberately NOT declared on this pack's manifest entry while the key is unset, because keyBlockedTools() in the gateway blocks every tool in a pack whose declared platform key has no secret — which would sink the two keyless catalogue tools that work today. The keyed tool declares _apiKey as required, so routing sinks only that one. Add "platformKeyEnv": "PLATFORM_CDS_KEY" to the pack-manifest entry in the SAME change that lands the secret.

Not the same as copernicus-dataspace

copernicus-dataspace is the EU Copernicus Data Space Ecosystem — the Sentinel satellite imagery catalogue (catalogue.dataspace.copernicus.eu). This pack is the Climate Data Store (cds.climate.copernicus.eu): modelled and satellite-derived climate variables. Same programme, different host, different question.

Data sources

• https://cds.climate.copernicus.eu/api/catalogue/v1/collections — STAC collection list. It returns the WHOLE catalogue in one response and ignores limit and q, so the keyword filter and paging in cds_collections are applied here, not upstream.
• https://cds.climate.copernicus.eu/api/catalogue/v1/collections/{id} — one STAC collection.
• https://cds.climate.copernicus.eu/api/retrieve/v1/processes/{id} — the OGC API process description. Keyless, and it is where the request form lives; the STAC record alone does not say what a retrieval accepts.
• https://cds.climate.copernicus.eu/api/retrieve/v1/processes/{id}/execute — job submission. Auth is the PRIVATE-TOKEN header, not a Bearer token and not a query parameter. Unauthenticated it answers 401 "authentication required"; with a bad token, 401 "Authentication failed" — the two are distinguishable.

Accepting the dataset licence on the CDS website is a separate gate from holding a token: a valid token still gets a 403 for a dataset whose terms the account has not accepted.

Quick Start

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

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

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/copernicus-cds/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 1683+ 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/cds_collections \
  -H 'Content-Type: application/json' \
  -d '{"query":"sea ice","limit":5}'

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

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

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

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-copernicus-cds

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

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 Copernicus Cds data" })

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

More

• Docs and guides
• pipeworx.io

License

MIT