Standards

archAIc is producer-stance for every standard it speaks. The engine is the one-stop server that emits IIIF manifests, Linked Art records, ISAD(G) JSON, EAD, EAC-CPF, METS, PREMIS, ALTO, PAGE, hOCR, TSDO, AtoM CSV, OAIS DIP-zip, and a handful of cross-cutting surfaces (OpenAPI, MCP, webhooks, Change Discovery, OpenTelemetry). It does not consume AtoM, Archivematica, or Mnemian as a downstream client; sync flows outward and standards contributions are made from the producer side.

The list below groups every format archAIc emits by the level it describes: a page, a Document Logical Unit (DLU), a libro (whole volume), or the cross-cutting and compliance layers that wrap every endpoint.

Per page (folio level)

One file per scanned folio. Token-level transcripts and visual anchors derived from the same PrIx data.

• TSDO — JSON. archAIc-native page model, lossless source-of-truth.
• ALTO — XML, schema 4.4 (3.0 on request). Preservation and library ecosystem.
• hOCR — HTML. Older web-OCR ecosystem and Internet Archive BookReader.
• PAGE — XML, PRImA 2013-07-15. HTR research community, Transkribus, READ-COOP.
• CSV — flat per-line export for spreadsheet tooling.
• AnnotationPage — IIIF / JSON-LD. Transcription overlay rendered by Mirador and Universal Viewer.
• Per-page PDF — searchable PDF.

Reach: GET /v1/pages/{collection}/{libro}/{page}/export?format=…. See the IIIF page for the AnnotationPage variant.

Per DLU (logical document level)

Typed entities, events, relations, anchored to source. Catalog-grade records, addressable by DLU id.

• Linked Art — JSON-LD profile of CIDOC-CRM. See Linked Art and CIDOC-CRM.
• ISAD-JSON — JSON serialisation of ISAD(G) 26-element record.
• EAD — Encoded Archival Description 2002 XML.
• EAC-CPF — authority records for persons / corporate bodies / families.
• METS — Metadata Encoding and Transmission Standard XML.
• PREMIS — preservation provenance XML.

Reach: GET /v1/dlus/{id}/export?format=… and PATCH on the writable formats.

Per libro (whole volume level)

Volume manifests, harvest-ready for any IIIF or archival platform.

• IIIF Manifest 3.0 — JSON-LD. The primary handle for a volume.
• navPlace extension — IIIF geospatial extension for places mentioned in the libro.
• Content Search 1.0 / 2.0 — in-viewer search and autocomplete.
• AtoM CSV — bulk import into Access to Memory.
• DIP-zip → OAIS — Dissemination Information Package, METS-wrapped, PREMIS-provenanced.

The IIIF surface is the IIIF page. AtoM CSV and the bulk-export endpoints are on the AtoM sync page.

Cross-cutting and APIs

Discoverable, signed, agent-callable across every endpoint.

• OpenAPI — machine-discoverable API surface.
• MCP — search, browse, extract, cite, with OAuth 2.1. See MCP.
• Webhooks — HMAC-signed event delivery.
• Change Discovery — ActivityStreams 2.0 + IIIF Change Discovery 1.0. Compatible with Europeana / DPLA harvesters.
• Content State 1.0 — deep-link any viewer state.
• OpenTelemetry — metrics, logs, traces.

Compliance and governance

Human-in-the-loop by construction; air-gappable on request.

• CIDOC-CRM — full ontology, authorities as first-class citizens.
• ISAD(G) — DLU metadata mapped 1:1.
• EU AI Act — Articles 12, 13, 14: logging, transparency, oversight.
• GDPR — air-gappable, Article 15 / 17 actuator.
• WAI-ARIA — screen-reader and keyboard-navigable UI.
• SAML 2.0 + OIDC — SSO, per-resource visibility, quotas.

Section pages

• IIIF — Presentation 3.0, Image API, Content Search, navPlace, Content State, AnnotationCollection.
• Linked Art and CIDOC-CRM — the JSON-LD profile for the DLU graph.
• AtoM sync — REST push, smart bulk sync, delete propagation, bulk CSV / EAD / EAC-CPF.
• MCP — Model Context Protocol server, agent surface, the four-verb pattern.

Roadmap

Specs drafted but not yet implemented in the engine. Dates appear as they ship.

• RiC-O — ICA Records in Contexts 1.0 with the proposed archaic:Mention extension for line-level provenance. Lined up with the Artefactual AtoM-import pilot. Ticket #222 (spec_ric_o_export).
• IIIF Auth API 2.0 — three-endpoint facade over the customer’s SSO session. Gated on the SSO spec landing.
• Webhooks — full multi-subscriber fan-out for dlu.confirmed and related events. A single-target producer for AtoM sync already exists; the generic subscriber registry is the remaining work.