You can make your SDMX data discoverable via an MCP server hosted on our cloudflare platform. Some of the challenges of creating an MCP server include:
How do i help users discover the data that is useful?
We improve the discovery experience for users by using a knoweldge graph created from the SDMX data. For example, if somebody is trying to understand the drivers of GDP, we can use the knowledge graph to prompt users with related concepts and data relevant to those drivers (such as data on increasing role of technology in business productivty) that wouldn't be discovered by a simple search.
We know what they say, but what do they mean?
Many non-technical people can start with a question that doesn't use standard terminology, such as "I want to know how well-off we are as a country", or misuse existing terms like "income". The process of linking these terms to the underlying concepts is termed 'entity resolution', and we help users address the underlying question using entity resolution embedded in our MCP server.
How do we make it as easy as possible for people to use the MCP server?
By using a remote MCP server, setup is easy and has low overhead.
What does an MCP Server for SDMX look like?
We have architected MCP server to address these needs of data discoverability & understanding meaning and also perfomance and security.
Within our architecture we have embedded:
Add your details for a link to the MCP server we have set up linking the Australian Bureau of Statistics for demonstration purposes.