API and bulk data
Everything here is readable by machine and free to query. No key, no registration, no rate limit beyond what keeps the server up. If you are checking a claim made on this site, take the rows and check them.
Bulk export
Up to 5,000 rows per request, paged with
offset. Every row carries its own source,
source_url and rights_statement, so an exported
file can be judged without coming back here.
# everything, first page
curl "https://egyptianinscriptions.org/api/export.csv?limit=1000"
# only photographed objects from the Metropolitan Museum
curl "https://egyptianinscriptions.org/api/export.csv?source=MET&has_image=true"
# every ostracon carrying text, as JSON
curl "https://egyptianinscriptions.org/api/export.json?form=ostracon&has_text=true"
# page through: the JSON response returns next_offset until it is null
curl "https://egyptianinscriptions.org/api/export.json?limit=5000&offset=5000"
Filters: source (unit code, see Sources),
form (see the Browse facets),
has_image, has_text, limit, offset.
Single record
The full record, including inscriptions, cross-references, connected
entities and a _provenance note stating which parts are
catalogued and which are inferred.
curl "https://egyptianinscriptions.org/api/object/22"Reading the data honestly
These fields exist so you do not have to take a row on trust:
find_point_precision— approximate means the coordinate is a site centroid, not the findspot. Nearly all geography in this corpus is at that level. inherited means the coordinate came from another record of the same text.has_inherited_fields— true when a findspot, site or date on this row was read from a sibling record across the Trismegistos spine rather than stated by this row's own source. The object page names the donor.match_methodon cross-references — anything other than explicit-source-field was matched by us, not asserted by the source.- Entities with
link_source: ai-visioncome from a machine reading of a photograph and can be wrong.
Licence and citation
The aggregation, cross-referencing and geocoding are free to reuse. The
underlying records belong to the institutions listed on
Sources and carry their own terms; each row's
rights_statement and source_url point at them.
Check those before republishing images or text.
Cite the database as:
Egyptian Inscriptions Database. https://egyptianinscriptions.org
Accessed <date>. 193,867 records from 25 sources.
Cite an individual record by its permanent URL,
https://egyptianinscriptions.org/object/<id>, and where possible cite the holding
institution's own record too — every page links to it.
Something wrong?
Every object page has a correction form, and corrections are read. If you need a cut of the data the export cannot express, or a full dump, say so through the contact details on the About page.