Statcan

What Statcan MCP server does

The Statcan MCP server exposes Statistics Canada's official statistical data through MCP. Its coverage includes headline Canadian indicators such as consumer price inflation, unemployment, GDP, and quarterly population estimates for Canada and its provinces and territories. It also provides access to the broader Statistics Canada cube catalogue, including areas such as trade, agriculture, and other published series.

Two access patterns are available. Indicator and vector tools offer a direct route to common measures or a known numeric series ID. Cube tools support discovery: list tables, inspect dimensions and members, identify vectors, then request observations for a selected coordinate. A CSV tool returns a Statistics Canada download URL without downloading the file itself.

How it works

The Statcan MCP server uses Statistics Canada's free, keyless Web Data Service. No API key or authentication header is required. The hosted MCP endpoint is available at https://gateway.pipeworx.io/statcan/mcp; the full Pipeworx endpoint is also supported. A local package can communicate over stdin and stdout with npx -y @pipeworx/mcp-statcan.

For cube queries, statcan_cube_metadata describes the dimensions and members of a product. statcan_cube_vectors can then enumerate vector IDs, coordinates, and member labels, with optional filters. Use the resulting coordinate with statcan_cube_data for the latest requested number of observations. Coordinates are dot-separated positions, and unused trailing dimensions require zeros.

The product ID may be supplied as an eight-digit WDS identifier or as a public table number such as 32-10-0121-01. Large cubes should be filtered before enumeration because unfiltered dimensions can produce a large combination set.

Setup and configuration

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

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

For a standalone local connection, use:

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

The gateway endpoint includes shared Pipeworx meta-tools in addition to the StatCan tools. The standalone process exposes only this package's tools. The data service itself is keyless, and no environment variables are required by the supplied setup examples.

Tools and capabilities

• statcan_indicator: Retrieves common Canadian indicators by name and geography.
• statcan_series: Fetches observations for a numeric Statistics Canada vector ID.
• statcan_list_cubes: Lists available cubes with product IDs, titles, CANSIM IDs, dimensions, and release dates.
• statcan_cube_metadata: Returns cube dimensions, members, frequency, geography, and release information.
• statcan_cube_vectors: Maps filtered cube members to vector IDs and coordinates.
• statcan_cube_data: Returns recent observations for one cube coordinate.
• statcan_table: Returns rows from a curated set of cached high-demand tables, with paging and filters such as geography and reference date.
• statcan_changed_series: Finds series changed on a specified date, defaulting to the current date.
• statcan_csv_url: Produces a direct full-cube CSV URL.

Limitations and notes

The Statcan MCP server is an independent, unofficial integration and is not affiliated with Statistics Canada. statcan_table covers only a curated, refreshed set of tables; use live cube queries or the CSV URL tool for tables outside that set. statcan_cube_data handles one coordinate per call, so broad comparisons may require vector enumeration and pagination. Series marked terminated have stopped publishing, and a cube's overall date range does not necessarily describe every series. Statistics Canada data is covered by its Open Licence and requires appropriate source acknowledgment.