Unauthenticated DCAT catalog search API

An unauthenticated catalog search API is a search endpoint that needs no key, no registration and no login - you issue GET requests and parse JSON. Data.gov's is the cataloged example: documented at resources.data.gov/catalog-api with GET /search as the primary method.

The cataloged industry research describes the query surface directly: 'keyword, organization-type and bounding-box searches over Data.gov's 552,271 records ... pulled programmatically through the unauthenticated DCAT search API', and its FAQ states that 'a public unauthenticated JSON API at resources.data.gov/catalog-api serves GET /search responses shaped as DCAT records'. The same portal record details the endpoint family - GET /search with q, per_page, org_id, org_type, keyword, after, spatial_filter and sort parameters, plus GET /harvest_record/<uuid> with /raw and /transformed DCAT variants and GET /api/keywords. The housewares-specialties brief counts the access split the same way, listing 'Data.gov's unauthenticated DCAT search' among the three official APIs serving its five pooled datasets.

Keyless access removes the onboarding step but also removes the usage contract, which changes how you build on the endpoint.

• No key means no quota to negotiate. The cataloged record describes no authentication and no published rate limit, so polite client behaviour is your only throttle - cache responses and paginate rather than hammering /search.

• Discovery is free; the data still is not guaranteed. The 552,271 records are metadata; distributions live on agency portals with their own licences, so a search hit does not settle reuse rights.

• Bounding-box search changes the workflow. Because spatial_filter is a documented parameter, regional harvesting is a query parameter instead of a post-filter - a real difference for state-by-state collection.

• Compare against the keyed alternative. GSA's v4 catalog API at api.gova.gov/technology/datagov/v4/search requires an api.data.gov key with 1,000 requests per hour for personal keys; the unauthenticated endpoint trades that ceiling for zero setup.

1. Verify the endpoint needs no key. The cataloged FAQ describes 'a public unauthenticated JSON API at resources.data.gov/catalog-api'.
2. Check the filter set before writing code. Documented parameters include q, per_page, org_id, org_type, keyword, after, spatial_filter and sort; confirm the ones your pipeline needs are actually supported.
3. Test the bounding-box filter on a known geography. The cataloged use case runs 'keyword, organization-type and bounding-box searches over Data.gov's 552,271 records' - validate spatial results against a region you can count by hand.

The record shape this API returns, and the industry pages using it:

• dcat-distribution-fields — The per-file fields (downloadURL, mediaType, licence, byte_size) each search result carries