Exporting Your Results

Exporting is how Acervo's outputs leave the system. This section covers the two export granularities (whole batch or single document), the available formats, which one fits which downstream system, the configurable ZIP for bundling multiple formats, and the dLOC submission handoff. It also covers reprocessing. Always user-initiated, since Acervo deletes source documents after each run. For direct sends to ArchivesSpace, see §7 Delivering & the ArchivesSpace Integration.

The sample folder does not export in any format; see §7.

Batch-level vs document-level exports

You can export at two granularities

• Batch-level export · Download every document in a batch as a single package. Use this when the whole batch is ready to ship to your downstream system, or when you'd rather grab everything in one download and pick out the records you need from the export afterwards (there's no penalty for ignoring records in a downloaded file). Open the batch from your sidebar's Recent list or from the History page (see §5 Reviewing in the Workbench).
• Document-level export · Export a single document's record. Navigate into the batch and select an individual document from the sidebar. Use this when you want a focused file for one record at a time.

Within either granularity, the format selection works the same way. Pick a single format, or bundle multiple formats together as a configurable ZIP (see The configurable ZIP below). The ZIP configures which formats are included, not which documents. Document selection is determined by your granularity choice above.

Available formats

Acervo exports in seven formats. Each is generated from the same underlying record, so you can pick whichever (or several) fit your downstream needs.

• CSV · Spreadsheet-friendly. A flat tabular view of the entities, subject headings, and core metadata. Good for spot-checks, bulk review, or importing into systems that accept CSV.
• dLOC CSV · A specialized variant of CSV shaped for the Digital Library of the Caribbean's ingestion.
• Dublin Core (XML) · Interoperable metadata for OAI-PMH harvesting, institutional repositories, and cross-collection discovery.
• EAD3 (XML) · Modern Encoded Archival Description for finding aids. Includes controlled access terms with authority URIs and DACS-compliant descriptive elements.
• EAD2002 (XML) · Older EAD revision. Needed if your downstream finding-aid system doesn't yet accept EAD3.
• MODS (XML) · Library-standard bibliographic descriptions with typed subjects, authority URIs, and multilingual metadata.
• OCR text · Raw OCR output without metadata structure. Useful for full-text search, indexing, or any workflow that needs the readable text directly.

Every export carries your repository identity from the institution profile (publisher, repository code, country code, language of description) plus the collection, rights, and donor facts declared with the batch. See §3 Uploading Documents for how the two fit together.

When to use which format

A practical mapping of formats to common destinations

• Archival finding-aid system · EAD3 if it's accepted; EAD2002 otherwise
• Library catalog or ILS · MODS or CSV (whichever your import tool accepts)
• Institutional repository or digital asset management system · Dublin Core or CSV
• Digital Library of the Caribbean (dLOC) · dLOC CSV
• OAI-PMH harvest · Dublin Core
• Spreadsheet review or hand-curated import · CSV
• Full-text search index · OCR text
• Mixed handoff to multiple systems · bundled ZIP (see below)

The configurable ZIP

When a single format isn't enough, you can bundle several together as a single ZIP file. Pick which formats to include. Useful when you need a CSV for spreadsheet review and an EAD3 for finding-aid ingestion, or when you're handing off material to multiple downstream systems and want one delivery package. The ZIP container is fully configurable. You choose the formats that go in, and Acervo bundles them together for download.

The ZIP configures which formats are included, not which documents. Document scope comes from your granularity choice. See Batch-level vs document-level exports above. If you want a slice of documents from a batch, you can either export at the document level individually for each, or just download the whole batch and use what you need from the resulting file.

Every ZIP also carries a manifest, in a readable text file and a machine-readable JSON file, listing what was exported and when, the formats included, the pipeline version, a SHA-256 fingerprint for every file, and per-document page counts, entity counts, and confidence summaries. It is the receipt for the handoff. A colleague or a downstream system can verify the package without opening every file.

The Digital Library of the Caribbean (dLOC) handoff

For dLOC partners, Acervo generates a dLOC-ready CSV that you submit to dLOC for ingestion.

Acervo does not ingest on your behalf. The handoff stays under your control

• You decide when to push to dLOC and what to include
• You handle dLOC's submission process
• You stay in dialogue with dLOC about ingestion outcomes

This is intentional. dLOC ingestion has its own review and curation, and your institution remains the responsible party for what enters the consortium's collections.

One more look before production

Your curated work is already in the export, edited titles and descriptions, corrected identifiers, chosen labels and selection included. If you still want to refine further, two places work.

• In the exported file itself · open it in your tools of choice (a spreadsheet for CSV, an XML editor for EAD/MODS/DC, a MARC editor if you're converting). Make corrections, save, then proceed to import.
• After import · bring the export into your downstream system (ILS, DAM, finding-aid system, repository) and make corrections there using the system's native editing tools.

Either path works. Pick whichever fits your team's workflow. The principle is the same. Don't treat the export as final the moment it lands. Even after a thorough workbench review, the export deserves one more look before it goes into production.

Reprocessing

Reprocessing is always user-initiated. Acervo never reruns a batch on its own. You'd reprocess to

• Apply a newer model that's improved since your original run
• Pick up an updated vocabulary (FAST/OCLC, LCSH)
• Try a different model choice (e.g., escalate from Standard to Premium for difficult pages)

Reprocessing requires re-uploading the source documents. Because Acervo deletes the originals after each processing run (see §10 Privacy & Data Handling), the source has to come from your archive each time. The result is a fresh batch with updated outputs; your previous batches stay in your history for reference.

If your goal is just to re-export an existing batch in a different format, you don't need to reprocess. Return to the batch from your sidebar's Recent list or the History page (see §5 Reviewing in the Workbench) and pick a new format from there.

Where to go next

• §9 Credits & Billing · How processing charges work, including for reprocessing runs
• §10 Privacy & Data Handling · What Acervo retains vs. what's deleted