Dbnomics

What the Dbnomics MCP server does

The Dbnomics MCP server gives an MCP-compatible agent access to data published through DBnomics. DBnomics aggregates statistics from more than 80 providers, including Eurostat, the U.S. Bureau of Labor Statistics, the European Central Bank, FRED, the IMF, the OECD, the World Bank, and national statistical agencies.

The server is intended for questions that require structured public statistics rather than general web search. An agent can locate a relevant provider or dataset, identify a series, and request its observations. Provider identifiers are uppercase, such as ECB, BLS, EUROSTAT, IMF, OECD, WB, and FRED; dataset and series identifiers depend on the provider.

How it works

The implementation uses the public DBnomics JSON API at https://api.db.nomics.world/v22/. MCP clients call the server's tools, and the server retrieves the corresponding data from that API. Authentication is not required.

There are two connection modes. The hosted option uses a Pipeworx gateway URL and can also expose shared gateway tools, including ask_pipeworx, discovery, search, and memory-related tools. Those shared tools are added by the gateway and are not part of the standalone package. The local option communicates over stdin and stdout and exposes only the Dbnomics tool set.

The hosted endpoint can also be called over HTTP without an MCP client. For example, a client can send a JSON POST request to the gateway's dbnomics_search endpoint.

Setup and configuration

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

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

For local use, configure the client to run the package with npx:

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

Running npx -y @pipeworx/mcp-dbnomics directly can confirm that the local process starts. The standalone mode does not require a Pipeworx account or a gateway round trip. No environment variables are documented for setup.

Tools and capabilities

The Dbnomics MCP server provides these data operations:

• search(query, limit?, offset?) performs full-text searches across providers.
• list_providers() returns the available data sources.
• list_datasets(provider, limit?, offset?) lists datasets belonging to a provider.
• get_series(provider, dataset, series_code, observations?) retrieves a selected time series and, optionally, its observations.
• find_series(provider?, dataset?, dimensions?, limit?, offset?) filters series using structured criteria.

The gateway connection also supports ask_pipeworx, which lets an agent submit a natural-language question for routing to an appropriate tool. The gateway's tools/list response may therefore contain substantially more tools than the five Dbnomics operations listed above.

Limitations and notes

Dataset and series codes are provider-specific, so an agent may need to search before it can request a series. The hosted pack endpoint lists shared Pipeworx meta-tools in addition to the pack's own tools; the local package does not. The hosted and full-gateway URLs use the same gateway and data sources, but they differ in which pack tools are listed directly. The project is released under the MIT license.