What DSpace is, and why institutions choose it
DSpace is open-source software for building institutional repositories: the platform where universities, research centers and public bodies publish, preserve and give open access to theses, articles, research data and reports. It was created by MIT Libraries and HP Labs, first released in 2002, and is maintained today by an international community with LYRASIS as its organizational home. According to the DSpace program, it powered more than 3,200 known installations worldwide in 2024, and an independent study of 2,729 institutional repositories found DSpace behind 42.3% of them, the largest share of any platform.
Modern DSpace is two applications: a Java backend that exposes a REST API, and an Angular frontend that users see. Underneath sit PostgreSQL for metadata, users and permissions, Apache Solr for search and usage statistics, and the assetstore that holds the deposited files. The result is a standards-based platform, with Dublin Core metadata, OAI-PMH harvesting, SWORD deposit and Handle persistent identifiers, but also one with more moving parts than a typical web application.
DSpace is not an integrated library system. Your catalog and loans belong in an ILS such as Koha; the repository publishes what your institution produces. Most universities need both, and they work best when planned together. Our guide to the institutional repository explains the difference.
How BiblioLabTech implements DSpace
BiblioLabTech delivers DSpace projects end to end, from the first assessment to the last upgrade, so your library deals with one team throughout.
The information model comes first
Before installing anything, we agree the structure of communities and collections, item types, submission forms, review workflows, access policies, embargoes and deposit licenses. These are the decisions that make or break a repository, and changing them with 10,000 items inside costs far more than settling them in a meeting.
Migration that keeps every citation working
A repository migration moves four layers: descriptive metadata, files, the community and collection structure, and persistent identifiers. The last layer is the one that cannot break, because Handles and DOIs are already cited in papers, theses and CVs. We decide at the start whether to transfer the Handle prefix or keep old identifiers with redirects, load a pilot batch of 100 to 500 representative items in a mirror environment, reconcile counts of items, files and sizes, and verify that every identifier resolves after cutover. Embargoed items are always part of the pilot, because an embargo that does not travel is the error with legal consequences. Read more in our guide to DSpace migration and upgrades.
Upgrades follow the same discipline. DSpace supports its three most recent major releases: since DSpace 10.0 shipped in May 2026, the supported lines are 8.x, 9.x and 10.x, and 7.x is end of life. Themes built for the old XMLUI or JSPUI interfaces are rebuilt in Angular, not migrated, and usage statistics stored in Solr are exported explicitly so they are not lost.
Visibility in aggregators and search engines
A repository that harvesters cannot reach exists only for people who already know its address. We configure the metadata Google Scholar expects, expose OAI-PMH with contexts that follow OpenAIRE guidelines, and register the repository in the relevant directories and aggregators. The repository lives on your own domain, never the provider's, so every URL you publish stays yours.
Training and support
Library staff who manage collections and review metadata, depositors in faculties, and IT each receive their own training on your real instance, and every session is recorded. After launch you move to annual support with agreed response times and planned upgrades, tested before they reach production.
DSpace-CRIS: from repository to research information system
A standard repository answers the question "where is this document". DSpace-CRIS also answers "what did this researcher, unit or funded project produce", because it models researchers, projects, organizational units and funding as connected entities instead of text inside a metadata field. On top of those entities it builds researcher profiles, collaboration networks, h-index and citation counts per researcher and unit, and CVs generated from data that already exists.
Its value comes from global identifiers: ORCID for researchers, ROR for organizations, DataCite for DOIs on datasets and other outputs, and OpenAlex to enrich and cross-check what the institution reports. Crossref, PubMed and arXiv can be used as open sources, and Scopus and Web of Science can be connected where the institution holds those subscriptions.
The timing matters. In November 2025, DSpace governance approved merging DSpace and DSpace-CRIS into a single platform. DSpace 10.0 is the first merged release and the process is set to conclude with DSpace 11.0, so starting with standard DSpace today does not close the door to CRIS. BiblioLabTech recommends filling the repository first and adding the CRIS layer as a second phase with its own budget, while making the right modeling decisions (normalized author names, ORCID, ROR) from day one.
AI metadata extraction, included
Many new repositories launch nearly empty, because describing years of accumulated theses by hand never ends. Every DSpace implementation by BiblioLabTech includes MetaExtract API at no extra cost. It reads each PDF and pre-fills the deposit form with title, authors, abstract, keywords and your local fields; the librarian reviews and approves instead of retyping. That is what makes it feasible to load a large historical collection without hiring extra staff for data entry.
Hosting: your servers or ours
DSpace has no license, but it needs a server running around the clock and someone to maintain it. With managed hosting, BiblioLabTech runs the server with HTTPS, monitoring, security updates and backups that cover the two things a repository backup must include: the PostgreSQL database and the assetstore with the files. Backing up only the database leaves a repository that shows every record and cannot deliver a single file. The domain, the certificate and the Handle prefix stay in your institution's name. Institutions whose policies require data to stay on campus keep the server, and we administer DSpace remotely. Compare both models in our guide to DSpace hosting.
What drives the cost of a DSpace project
DSpace and DSpace-CRIS are free, so the cost of a project is the cost of the services. In BiblioLabTech's university projects, a new single-campus repository without the CRIS layer typically costs in the order of USD 3,000 to 5,000 in the first year, covering implementation, initial load, training, support and hosting; from the second year, the recurring cost drops to hosting and support. The price moves with the volume of files in gigabytes rather than the number of items, the quality of the source metadata, integrations, frontend customization and the CRIS layer. The guide to DSpace implementation cost breaks it down component by component.
Where DSpace with BiblioLabTech is running today
- Universidad Americana (Paraguay) runs its institutional repository on DSpace hosted and operated by BiblioLabTech in the cloud, with MetaExtract API in the deposit workflow. The contract has been renewed and the university is evaluating an expansion to DSpace-CRIS. Read the case study.
- Escuela Colombiana de Ingeniería Julio Garavito (Colombia) runs its DSpace repository alongside its Koha catalog, the BiblioMeta discovery layer and a reference chatbot, on a dedicated university server administered by BiblioLabTech. See the full case study.
- Universidad Dr. Andrés Bello (El Salvador): BiblioLabTech is implementing NexoHub, the university's research information system on DSpace-CRIS, integrated with ORCID, ROR, OpenAlex and DataCite.
- Universidad Católica Luis Amigó (Colombia) has run MetaExtract API in production on its DSpace repository since October 2024. See the case study.
BiblioLabTech works remotely with institutions in any country. Our team is based in Colombia (UTC−5), which overlaps the full working day across the Americas and the afternoon in Europe and Africa. Tell us about your repository and we will reply with a scoped proposal, quoted with and without hosting.
DSpace is free software developed by its community, with LYRASIS as its organizational home, at dspace.org.