JSTOR has launched transcript generation within JSTOR Seeklight, expanding its AI-assisted collection processing capabilities beyond metadata to include full-text transcription of handwritten, typed, and mixed-media documents. The new feature is now available to all Tier 3 charter participants in JSTOR Digital Stewardship Services.

Developed in collaboration with librarians and archivists, this new functionality enhances discoverability, supports compliance with accessibility requirements, and maximizes the research value of digital collections. Transcripts can be generated during upload or on demand, reviewed and refined in a built-in editor, or replaced with institutional versions—ensuring expert oversight by archivists and librarians, along with institutional control.

“In our conversations with librarians and archivists, transcripts have consistently emerged as essential to ensuring accessibility and advancing discoverability,” said Roger Schonfeld, managing director of JSTOR Digital Stewardship Services. “From our experience with journals and books on JSTOR, we know that metadata is important, but full text is what drives transformative usage and research impact. By enabling JSTOR Stewardship participants to make difficult-to-process materials machine-readable—and by including this capability as part of the core Tier 3 offering—we’re advancing our shared mission to expand access to knowledge and education.”

JSTOR Seeklight, which launched with metadata generation capabilities in April of this year, was honored with the 2025 C.F.W. Coker Award from the Society of American Archivists this summer for innovation in archival description.

Institutions interested in exploring AI-assisted digital stewardship can join the JSTOR Digital Stewardship Services charter program. Charter participants gain early access to new capabilities like transcript generation, work directly with JSTOR and peer institutions to refine tools and practices, and take part in a collaborative community shaping the responsible use of AI in libraries and archives.