When to use this registry

Use dataportals-registry when you need stable metadata about data catalogs — not the datasets inside those catalogs.

Use this repository for

Finding portals, geoportals, repositories, metadata catalogs, and related infrastructure by country, type, or software
Joining catalogs to Wikidata, re3data, FAIRsharing, and similar identifiers
Landscape analysis (how many CKAN sites in France, which geoportals expose CSW, …)
Training or evaluating catalog-metadata classifiers, with the biases in DATASHEET.md
Feeding a downstream search index that crawls endpoints listed on each record

Querying dataset records inside a portal (CKAN packages, Dataverse studies, STAC items)
Production search APIs or MCP servers — this repository is reference data only. Search: dateno-api. Catalog browse UI: dataportals-web. Harvesting catalog contents: reaper.
Real-time uptime SLAs — liveness probes are report-only
Exhaustive coverage of every country; the United States is heavily over-represented

“Is there an open data portal for city X / country Y?”
Use this registry: filter catalog_type and coverage. If nothing matches, follow discovery.md (search engines: discovery-search-tools.md; platforms: opendata, geoportals, scientific, metadata, indicators, other types) before adding a record.
“Which catalogs run CKAN / GeoNetwork / Dataverse?”
Use this registry: filter software.id in DuckDB / Parquet (query-examples.md).
“Give me the API URL to harvest.”
Use this registry: read endpoints[] and api_status. Then harvest the remote catalog (harvest.md). Scientific IRs: harvest-scientific.md. Open data: harvest-opendata.md. Geo: harvest-geoportals.md. Indicators: harvest-indicators.md. Metadata: harvest-metadata.md. Aggregators/ML/custom: harvest-other.md. Shared protocols: harvest-protocols.md. Incremental: harvest-incremental.md. EO: harvest-earthdata.md. Biodiversity: harvest-biodiversity.md. Viewers: harvest-viewers.md. Identifiers: harvest-identifiers.md. Output: harvest-output.md.
“Search for a specific dataset titled …”
Out of scope for this repo’s data. Harvest the catalog (harvest.md) or use a downstream search index.
“Is this URL currently up?”
Optional signal only: dataquality/liveness_report.jsonl. Status in YAML is curated, not a live probe.

Need	Where
Catalog metadata (this repo)	YAML + DuckDB/Parquet exports
Production catalog/dataset search API	dateno-api
Harvest datasets listed in catalogs	reaper (production); recipes: harvest.md
Public catalog website	dataportals-web
Country and organization reference data	internacia-db
Unified file reading/writing	iterabledata