The IIIF Endpoints

Every public text and edition is served as IIIF Presentation 3.0 and W3C Web Annotation, live from the database. There is no export step and no separate API key: the documents below are the publication.

For what the layers mean, see Annotations and IIIF and Deep-Zoom Images. This page is the URL table.

Base

https://<supabase-host>/functions/v1/iiif/v3

All endpoints are GET, return JSON-LD, and are CORS-open. <supabase-host> is the backend for the installation you are using — take it from the iiif-content parameter of any Viewer link, which names a manifest on the same base.

Entry points

Endpoint	Returns
/collections/{id}	Collection — the texts in a collection, in the order set in the Editor
/texts/{id}	Manifest — one text, its canvases, its metadata (including custom metadata), its editions
/editions/{id}	AnnotationCollection — one edition

A manifest is the normal place to start: it names everything else. It is also what the Viewer opens.

Canvases

Endpoint	Returns
/images/{id}/canvas	Canvas — one manuscript image, with its dimensions and metadata
/images/{id}/canvas/page	The canvas's AnnotationPage
/images/{id}/canvas/annotation	The painting annotation placing the image on the canvas

For an image served by the image server, the painting annotation carries an ImageService3 service at level2, which is what makes deep zoom possible. A plain image has no service — the body is the file itself.

Annotation pages

One page per layer per edition. Each is an AnnotationPage whose partOf names the edition's annotation collection.

Endpoint	Contains
/editions/{id}/segments	Every akṣara
/editions/{id}/lines	Every line
/editions/{id}/tokens	Every token
/editions/{id}/sequences	Every sequence
/editions/{id}/lemmas	Every lemma attested, with its attestations
/editions/{id}/inflections	Every inflection used, with its attestations

"Segment" is the akṣara

The API name for an akṣara is segment, throughout. See the glossary.

Individual annotations

Endpoint	Motivation	Target
/segments/{id}	describing	The canvas, with an SVG selector giving the outline
/lines/{id}	linking	Its segment annotations, in order
/tokens/{id}	linking	Its segment annotations, in order
/sequences/{id}	linking	Its token annotations and child sequences
/lemmas/{id}	linking	Every token annotation attesting it
/inflections/{id}	linking	Every token annotation carrying it

The layering is the point: only the segment annotation targets an image region. Everything else targets other annotations, so the editorial apparatus survives publication without being flattened.

Bodies are TextualBody in HTML — a segment's graphemes with their marks, an inflection's grammatical roles, a token's grapheme offset — so a generic annotation client can render them without knowing anything about Gāndhārī.

Visibility

Requests are resolved under the same rules as the applications:

• Anonymous — public texts and public editions only.
• With your credentials, sent as a bearer token in the Authorization header — additionally everything you own, published or not.

An unpublished edition returns nothing to an anonymous caller, and does not appear in its text's manifest. There is no unlisted-URL loophole. See Ownership and visibility.

Notes for reuse

The documents are generated on request, so they reflect the edition as it stands now. Editions have no version history; if you need a fixed citable state, keep the report.

Image rights are not edition rights. The annotations are the editor's scholarship; the photographs are governed by the holding institution's terms, which travel with the manifest. Publishing an edition licenses neither. See Using images outside Glycerine Aksara.