Japan Search Open Heritage Metadata - CC0/PDM, Per Object

Japan Search's openly-licensed heritage metadata as clean, per-object records - one record per object whose rights are Public Domain Mark (PDM) or CC0. Museum, library and archive object metadata - title, type, date, institution and rights - aggregated across Japan's cultural databases, with Japanese text delivered verbatim.

What one record represents

The source is Japan Search (jpsearch.go.jp), the national aggregator, via its jps-cross cross-search API. This cell is scoped to objects whose rights are Public Domain Mark or CC0. Each record is one object: its id, title (Japanese, and English where provided), category, object type (image, text, sound, ...), date text, location, the providing institution, the rights value, an access flag, a description and a link to the object at its source.

Coverage and volume

The live open-rights set is 945,789 objects - 855,083 under Public Domain Mark and 90,706 under CC0 (measured at build time from the API's rights facet). The Actor sweeps both rights values and delivers their union.

Sol's Wave-3 index put this door at 945,789 CC0/PDM records; measured live at build time the two rights facets sum to exactly 945,789 (855,083 PDM + 90,706 CC0) - matching the index.

The Actor pages the jps-cross API 100 objects at a time, sweeping the PDM slice and then the CC0 slice, and stops as soon as your Maximum records cap is met.

Licence and rights

This cell delivers only objects whose rights are PDM or CC0 - open, public-domain-equivalent terms. The rights filter (f-rights=pdm and f-rights=cc0) is baked into the query, not a buyer option, and a per-record assertion rejects any object whose rights are not PDM or CC0 (verified with a planted-row test). Japan Search API use is governed by the Japan Search policy:

Japan Search API use is governed by https://jpsearch.go.jp/en/policy. This cell delivers ONLY objects whose rights are Public Domain Mark (pdm) or CC0; the rights filter is baked into the query. Metadata is aggregated from contributing institutions, cited via the provider/database fields.

Rights apply to the object's contents as marked by the contributing institution; the rights, provider and database fields let you attribute and trace each object to its source.

Person-data policy

Japan Search objects can carry contributor, creator-person, donor and rights-holder-person fields. This Actor structurally excludes all of them: the contributor field and any creator/donor/rights-holder person data are never read. The institutional fields that are kept - provider, owner_org, database - are organisations, never persons, and a per-record assertion rejects any contributor/creator/donor/person field (verified with a planted-field test). Object titles and descriptions are historical/cataloguing text about the artefact, delivered verbatim. No natural-person data is processed.

Interpretation caveat

One record per Japan Search cross-search object whose rights are Public Domain Mark or CC0 (~945,789: 855,083 PDM + 90,706 CC0). Heritage object metadata: id, title (Japanese verbatim + English where provided), category, type, date text, location, providing institution, rights, description, link. Japanese text is delivered verbatim. Contributor, creator-person, donor and rights-holder-person fields are structurally excluded.

Values are reproduced verbatim from the API; the Actor never rewrites a field, and Japanese text is delivered as-is with no transliteration - relabelling would be a modification. The date_text is the source's own period/date string (often an era name), not a normalised date. A field the source leaves empty is delivered as null.

Data quality and freshness

The record_id is the object id, so the dataset is safe to diff, deduplicate or upsert. Every run re-reads the live API, so the data is as fresh as Japan Search aggregates, and each record's observed_at stamp dates the snapshot. The run's RUN_RECEIPT records the per-rights totals alongside how many records were delivered and charged.

Provenance and compliance

Every run reads jpsearch.go.jp/robots.txt at runtime; the gate result (URL, status, byte length, SHA-256 of the policy) is written to the run's RUN_RECEIPT, and the jps-cross path (outside the portal's disallowed /data/, /rdf/, /term/, /entity/ paths) is confirmed crawlable before any data request. The API is keyless. The Actor never bypasses a block or fetches through a mirror.

Inputs

• Maximum records (maxRecords) - hard cap on heritage-object records delivered and billed.

Output

Records land in the Actor's default dataset and export as JSON, CSV, Excel or via the Apify API. A tabular overview view surfaces object id, title, English title, type, date text, provider and rights.

Fields in detail

The record leads with object_id and title (plus title_en), then category, object_type, date_text, location, the institution fields provider/owner_org/database, the rights value, access and contents_access, description, link_url and thumbnail_url. The provenance block closes every record.

Typical uses

Archive, licensing and digital-collection teams use this cell to find commercially reusable heritage metadata - one queryable table of nearly a million PDM/CC0 objects across Japan's cultural institutions, filterable by type, provider or era, with a link back to each object at its source. Because the rights are baked to public-domain-equivalent terms, every row is safe to reuse, and the provider/database fields let you attribute per object. A scheduled run keeps a downstream heritage index current as Japan Search aggregates more collections.

Scaling and limits

Set Maximum records low to sample or high to pull toward the full ~945,000-object set. The Actor pages the API 100 objects at a time across the two rights slices and delivers incrementally, so memory stays flat and you are billed only for what is delivered. Re-running picks up newly aggregated open-rights objects, and each record's observed_at stamp dates the snapshot.

Cross-database aggregation

Japan Search is an aggregator: it federates metadata from hundreds of museum, library, archive and academic databases across Japan into one cross-search profile (jps-cross). The provider and database fields on every record tell you which contributing institution and source database the object came from, so you can attribute per object and, if you need the fullest record, follow link_url back to the source. This cell reads the normalised cross-search layer, which gives a consistent field set across every contributing database - one shape to load, regardless of which institution holds the object.

Working with the Japanese metadata

Titles, descriptions, date text and location are delivered exactly as the source holds them, which for the great majority of objects means Japanese. The Actor does not transliterate or translate - relabelling a value would be a modification, and the discipline of this fleet is verbatim reproduction - but it does carry the English title (title_en) wherever the source provides one. The date_text is typically an era-based period string (for example a Japanese period name plus a century) rather than an ISO date, faithful to how heritage cataloguers record provenance. If you need romanised or normalised values, do it downstream; the record gives you the authoritative source strings to work from. This verbatim discipline is what lets an archive or licensing team trust the record as evidence of what the institution actually published, rather than as a lossy paraphrase - the difference between metadata you can cite and metadata you have to re-check.

Sibling Actors

It sits beside the fleet-1 Japan press-release cell (news vs heritage metadata). It shares its engineering - the runtime robots gate, offset paging with the rights filter baked in, push-then-charge billing and verbatim-value discipline - with the fleet's other public-data records Actors, and stays distinct from the Japan press-release cell (news vs heritage metadata).