For Archives

Ethics & Consent

The use of cultural materials in artificial-intelligence training is contested, and rightly so. Akshara works only with public-domain and rights-cleared collections, and never re-licenses material without partner sign-off. Institutional veto is absolute and durable: if a stewarding body withdraws an item, work, or collection at any stage, it is removed from the corpus.

For Indigenous and culturally sensitive holdings, we follow the CARE Principles for Indigenous Data Governance and the Local Contexts Traditional Knowledge labels, and defer to the community protocols and ethical-review processes already established by the steward institution.

Archives should be partners (credited, compensated, and consulted) in how the record they have kept enters the next era of human knowledge. Not extracted from. Not bypassed.

The Partnership

You retain everything we capture. Master files, derivatives, OCR, descriptive metadata, IIIF manifests: yours, in perpetuity, for any use you choose, with no exclusivity. Akshara structures the digitized corpus and licenses it to AI laboratories for model training; that revenue funds the digitization itself, returns a share to the institution, and underwrites preservation work that would otherwise go unfunded.

• You provide Rights-cleared materials and the curatorial relationship.
• We provide Capture, processing, metadata, delivery, and the licensing infrastructure.
• The public gains Access to collections that have remained dark for decades.

Standards & Methodology

Capture meets FADGI 4-star and Metamorfoze imaging guidelines, the standards used by the Library of Congress, the Smithsonian, and the Dutch national heritage program. Equipment is selected per material: overhead capture for bound volumes, planetary scanners for fragile or oversized items, dedicated stations for microfilm conversion. OCR is reviewed for accuracy on representative samples and tuned per script, language, and historical period.

Per item, you receive:

• Preservation TIFFs, uncompressed, 600+ ppi for textual material
• JPEG 2000 access derivatives
• PDF/A reading copies
• ALTO XML for OCR with coordinates, plus plain-text transcripts
• Descriptive metadata in MODS or Dublin Core, to your specification
• IIIF Presentation API 3.0 manifests on request
• PREMIS preservation events recorded for every item

How a Partnership Works

I. Selection. We work with your team to identify public-domain or rights-cleared materials suitable for digitization, typically print, manuscript, or microfilm collections. Your rights-and-reproductions staff retains final say on every inclusion.

II. Capture. On-site digitization, or insured transit to our facility, depending on fragility, scale, and your institutional preference. Climate-controlled handling. Conservator review of any item flagged for treatment. Chain of custody documented from accession to return.

III. Delivery. You receive full-resolution preservation files, access derivatives, OCR, metadata, and IIIF manifests, formatted to your specifications and ingested at your pace. Yours, in perpetuity, with no exclusivity clause.

IV. Licensing. We structure the corpus and license it to AI laboratories under terms that preserve provenance. Your institution is credited as the archival source on every derived work. Revenue is shared per agreement. Veto remains with you at every stage.

Logistics

• Insured shipping where transit is required; on-site capture for collections that should not travel
• Conservator-reviewed handling of fragile, oversized, or unique material
• Background-checked operators; climate-monitored workspace
• Materials returned in original housing, within an agreed timeline
• Pilot timeline: typically six to twelve weeks from MOU to first deliverable