Skip to main content

Connect to STH

Semantic Treehouse offers several connecting options for integrating with other tools and systems, including an external API, catalog exports and LLM integration through an MCP server.

In case you are missing a connecting option, or have trouble connecting through the current ways, get in touch with us via the Semantic Treehouse Discord.

Semantic Treehouse API​

The Semantic Treehouse API is decribed in the API section, as an OpenAPI specification.

Catalog exports in RDF (DCAT+)​

STH provides catalog exports to facilitate interoperability and the discovery of compatible datasets. The catalog is expressed using a semantic web standard, where the basis is formed by DCAT. This machine-readable export describes metadata of the environment's specifications, versions and distributions and allows DCAT compatible tooling to discover these specifications. Read more about catalog exports.

Connecting your LLM client (experimental)​

STH offers an experimental, proof of concept MCP server. It lets you connect your own LLM client (an AI assistant such as Claude) to a Semantic Treehouse environment, so you can ask questions about the content of that environment in plain language.

Experimental

The MCP server is a proof of concept. It may change or behave unexpectedly, and it is not enabled on every environment. If it is not available on your environment, contact your administrator.

What you can do with it​

The assistant works with your own account: it sees exactly what you are allowed to see in Semantic Treehouse, and nothing more. It can only read, so it cannot change anything in your environment.

You can ask it, for example, to:

  • find a specification and summarise what it is about;
  • explain the structure of a message model, ontology, taxonomy or JSON schema;
  • look up the entries of a codelist, or check whether a certain code exists;
  • show how two specifications are mapped to each other;
  • list the open issues of a project, or summarise the discussion on a specific issue;
  • give an overview of your groups, their members and their meetings.

As with any AI assistant, answers can be incomplete or wrong. Check important results in Semantic Treehouse itself.

Adding it to Claude Desktop​

  1. Open Claude Desktop and go to Settings > Connectors.
  2. Choose Add custom connector.
  3. Enter a name, for example Semantic Treehouse, and as URL the address of your environment followed by /mcp, for example https://your-environment.semantic-treehouse.nl/mcp.
  4. Click Add, and then Connect.
  5. A browser window opens. Log in to Semantic Treehouse if you are not logged in yet, and give Claude access to your account.

Semantic Treehouse is now available in your conversations. Try it out with a question such as "Which specifications can I see in Semantic Treehouse?".

Other LLM clients that support remote MCP servers can be connected in a similar way, using the same URL.