Search Switzerland's national metadata catalogue for data resources.

The primary entry point: use this to find out who publishes data on a topic before looking for a specific dataset. Returns titles, publishers, themes and portal permalinks.

Known limitation (verified live 2026-07-21): the upstream search index covers Datasets only. Filtering by types=["Concept"] or types=["DataService"] returns zero results even though those entities exist — use list_concepts and list_data_services for those.

Args: query: Free-text search term, e.g. "Sonderpaedagogik" or "Bildung". An empty string returns the full index (over 1000 records) and is not recommended. language: Language for titles and descriptions. types: Restrict to resource types. Effectively only "Dataset" works. themes: Filter by theme identifiers. publishers: Filter by publisher identifiers. limit: Maximum records to return (1-200). Upstream ignores paging, so capping happens in this server.

List registered datasets, optionally filtered by publisher.

Unlike search_catalog, this endpoint paginates correctly and covers the complete register (roughly 2000 datasets as of July 2026), including records the search index misses.

Args: publisher_identifier: Publisher identifier from list_publishers. access_rights: e.g. "PUBLIC", "NON_PUBLIC", "RESTRICTED". language: Language for titles and descriptions. page: 1-based page number. page_size: Records per page (1-100).

Retrieve the full, aggregated metadata record for one dataset.

This is the aggregated detail tool: a single call returns the contact point, temporal and spatial coverage, documentation links and every distribution with its licence — so search_catalog → get_dataset answers «who publishes it, through which interface, under which licence» in two calls, without a separate distributions or contact lookup.

Args: dataset_id: UUID from search_catalog or list_datasets. language: Language for titles and descriptions.

Get the downloadable files and access URLs for a dataset.

This is the «where do I actually get the data» tool. Each distribution carries its own format, licence and download URL — licences differ between distributions of the same dataset, so always read the licence field before reusing the data. (get_dataset returns these same distributions alongside the rest of the record.)

Args: dataset_id: UUID from search_catalog or list_datasets. language: Language for titles and descriptions.

List machine interfaces (APIs) registered by Swiss public bodies.

The strategic payload of this server: the national register of official APIs, with endpoint URLs and OpenAPI specification links where the publisher supplied them. Use this to discover whether an interface already exists before building a scraper.

Args: publisher_identifier: Publisher identifier from list_publishers. language: Language for titles and descriptions. page: 1-based page number. page_size: Records per page (1-100).

Retrieve the full record for one registered API, including endpoints.

Args: data_service_id: UUID from list_data_services. language: Language for titles and descriptions.

List registered public services (administrative offerings for citizens).

Args: publisher_identifier: Publisher identifier from list_publishers. language: Language for titles and descriptions. page: 1-based page number. page_size: Records per page (1-100).

List harmonised concepts and code lists of the Swiss administration.

Concepts are the semantic backbone of interoperability: shared definitions and code lists that different bodies agree to use. Note that these are not reachable through search_catalog.

Args: publisher_identifier: Publisher identifier from list_publishers. language: Language for titles and descriptions. page: 1-based page number. page_size: Records per page (1-100).

Retrieve one concept definition, including its value type and version.

Args: concept_id: UUID from list_concepts. language: Language for titles and descriptions.

List the individual codes of a code-list concept.

Use this to resolve official code values — for example the canonical list of a classification used across several federal datasets. Only concepts with concept_type == "CodeList" return entries.

Args: concept_id: UUID of a code-list concept from list_concepts. language: Language for entry names. Required by the upstream API. page: 1-based page number. page_size: Records per page (1-200).

List the public bodies that publish into I14Y.

Returns publisher identifiers needed by the other tools, plus the Swiss UID where available — the UID is the join key to register-mcp (Zefix).

Args: identifier: Filter by exact publisher identifier. uid: Filter by Swiss UID, e.g. "CHE-123.456.789". language: Language for names. page: 1-based page number. page_size: Records per page (1-100).

List the catalogues that feed into I14Y.

Each catalogue represents one contributing organisation's data collection.

Args: language: Language for titles. page: 1-based page number. page_size: Records per page (1-100).

Check whether the I14Y API is reachable and which endpoints respond.

Always returns an evaluable status rather than an empty result, so an agent can distinguish «no data matched» from «the source is down».