Copernicus Cds

What Copernicus Cds does

Copernicus Cds connects MCP clients to the Copernicus Climate Data Store operated by C3S/ECMWF. It covers the catalogue of modelled and satellite-derived climate datasets, including data associated with ERA5, CMIP6 and CORDEX projections, essential climate variables, and seasonal forecasts.

The integration is separate from the Copernicus Data Space catalogue. Copernicus Cds targets climate variables at cds.climate.copernicus.eu; the other service targets Sentinel satellite imagery.

How it works

The hosted MCP endpoint is available at https://gateway.pipeworx.io/copernicus-cds/mcp. A client can also connect to the full Pipeworx gateway, where the same pack is available alongside other data sources. Pack endpoints include shared Pipeworx meta-tools in addition to the pack-specific tools; ask_pipeworx can route questions across the wider catalogue.

A local npm bridge is available as @pipeworx/mcp-copernicus-cds. It runs with npx, forwards requests to the hosted gateway, and requires network access. HTTP callers can invoke tools through the gateway's /v1/tools/{tool} endpoints.

Setup and configuration

For a hosted MCP connection, add the endpoint to a client configuration:

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

For the local bridge, use:

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

Anonymous local-bridge use is limited to 50 calls per day. Setting PIPEWORX_API_KEY raises the allowance to 200 calls per day. PIPEWORX_META_TOOLS=1 adds the shared gateway tools to the local bridge's listed tools. The bridge can use NODE_USE_ENV_PROXY=1 with Node versions that support the built-in environment proxy behavior.

Tools and capabilities

• cds_collections lists or locally filters the catalogue by keyword, with limit and offset controls.
• cds_collection returns a collection's STAC metadata and, optionally, its retrieval form. The form identifies accepted variables, years, levels, geographic area, and output formats.
• cds_submit_request starts an asynchronous retrieval and returns a job identifier. It requires _apiKey.
• cds_job_status reports whether a job is accepted, running, successful, or failed, along with timestamps.
• cds_job_result returns the download URL, file type, size, and expiry after success. In-progress jobs return ready:false and their status.

Limitations and notes

Catalogue operations are keyless, but retrieval operations require a free CDS Personal Access Token passed as _apiKey. The token must use the CDS API's PRIVATE-TOKEN authentication mechanism. Dataset licence acceptance is a separate requirement; a valid token can still receive a 403 until the account accepts the relevant terms.

The upstream collection-list endpoint returns the complete catalogue and ignores its own query and paging parameters. The integration applies keyword filtering and pagination after receiving that response. Retrievals are asynchronous, so callers should retain the job ID and poll status before requesting the result. This is an unofficial integration and is not affiliated with the upstream provider.