- 0open-source repos
- 0background workers
- 0open standards spoken
- 0universities running it
A library that understands its own content
ELVIRA is more than a catalog. A pipeline of AI workers extracts text, tables, figures and formulas from each publication, embeds them with multilingual models, and exposes the result through semantic search, an AI assistant and standards-based lending.
Read
Embedded PDF viewer with annotation layer, BibTeX / RIS citation export and page-range sharing. Works for born-digital PDFs and OCR-recovered scans.
Search
Multilingual semantic search over the full corpus. A Slovak query finds relevant passages in English, Czech and German literature. An AI assistant answers questions over library content with passage-level citations.
Lend
OPDS 1.2 and 2.0 distribution. End-to-end Readium LCP lending — borrow, reserve, renew, return — with EdRLab certification in progress.
Two universities, one codebase
ELVIRA runs at two Slovak universities today. The portal handles per-site theming and configuration; the catalog underneath is the same software.
Primary production deployment. ELVIRA runs STU-wide; FIIT STU is the home faculty where the project is developed (instance at elvira.fiit.stuba.sk).
KU RužomberokCatholic University in Ružomberok (Katolícka univerzita v Ružomberku)Second deployment site. Portal themed for KU, same catalog software underneath.
Run your own deployment
Container images, deployment manifests and the OpenAPI protocol are all public. License is MIT (BSD-3-Clause for the Readium fork). Pull requests welcome.