Anyone who has ever manually typed the title, authors, abstract, and keywords of a research paper into a repository submission form knows the pain. It’s tedious and error-prone. Thanks to our latest contribution to the DSpace open-source project, GROBID-powered metadata extraction is available to the entire DSpace community. When you start a submission and upload your PDF first, GROBID detects metadata like authors, title, abstract, keywords, identifiers, and more. The recognized metadata is presented to you in the submission form, where you can change them or add more information, before you accept the license and click the button to submit your newest publication. It’s that simple and it makes the submission experience dramatically smoother for researchers and librarians alike.

What Is GROBID?

GROBID — short for GeneRation Of BIbliographic Data — is an open-source machine learning engine that converts raw PDF documents into structured bibliographic metadata. Originally started in 2008 and focused on technical and scientific publications, GROBID parses a PDF and returns a rich XML representation of its contents: title, authors, affiliations, abstract, keywords, identifiers (DOI, ISBN, ISSN), publication date, journal or conference source, bibliographic references, and more — typically with 80–95% accuracy and in a matter of seconds.

From DSpace-CRIS to DSpace: The Contribution

GROBID integration was originally a feature of DSpace-CRIS, developed by 4Science. As part of the ongoing merger of DSpace-CRIS into DSpace, we ported this feature to the main DSpace codebase and contributed it back to the open-source community. The pull request — DSpace/DSpace#11852 — was merged on September 18, 2026 and will ship with DSpace 11.

During the port, we didn’t just copy the old code over. We modernized the implementation and added a ready-to-use Docker Compose file so that anyone can spin up a local GROBID instance alongside DSpace with minimal effort.

We’re proud to bring this feature from DSpace-CRIS to the broader DSpace community. It’s a concrete example of how the merger is delivering real value to every DSpace repository worldwide.