Adopting ELVIRA

ELVIRA already runs at more than one institution, and it’s built to travel. This page is the honest version of “could we run this?” — what you get, what you need, and where the edges are.

It’s open source, so you can stand it up entirely on your own. You don’t have to, though — the team supports faculties and universities that want a hand running it.

Who it’s a good fit for

• A university or faculty library that wants its publications searchable and readable in the browser, without buying into a closed platform.
• Anyone who needs standards-based lending — OPDS catalogs and Readium LCP — so existing e-readers work out of the box.
• Teams comfortable running containers and a bit of Python. This is research-grade open source, not shrink-wrapped software.

What you get, out of the box

• A multi-tenant catalog — one install can host several libraries, each with its own entries, permissions and theme.
• OPDS 1.2 and 2.0 feeds, so reading apps just work.
• The portal and the embedded reader, themable to your institution.
• The worker fleet — text extraction, OCR, figure captioning, equation and video processing, semantic search — added as you need them.
• Optional Readium LCP lending. (Official EdRLab certification is in progress; see the status report.)

What you’ll need to bring

• Infrastructure. Somewhere to run containers, a database and the broker. The AI workers want a GPU to be comfortable; the core catalog does not.
• Identity. The portal expects university SSO; wire it to your provider.
• A theme. The portal is one codebase themed per deployment — add yours alongside fiit and ku.
• Content. Import from Kramerius / WebDAV, or upload over the REST API.

How to start

1. Stand up a local catalog with the Quickstart — it’s a few Docker commands.
2. Read the Architecture overview so the moving parts make sense before you scale any of them out.
3. When you’re ready to talk specifics — theming, SSO, LCP, a real deployment — open a discussion or get in touch. We’d genuinely like to hear from you.

The catch, stated plainly

ELVIRA is developed by students and researchers and moves at that pace. The core is in daily production and dependable; the AI services are newer and evolve release to release. Adopt the core with confidence, and treat the worker fleet as capable, active research you can grow into.