HTR for Latin stone inscriptions: a use-case on the cluster
A digital epigraphy project as an example of what the cluster's GPU nodes can be used for — from a single training run to a systematic parameter search.
Shared infrastructure forDigital Humanities in Austria
We pool compute, development services, digitization, and open-source software across Austrian institutions — so DH projects that lack resources or skills can still get the job done. A CLARIAH-AT initiative.
What we do: More compute, easier access, and better support for digital humanities — democratizing infrastructure for all.
Resources
What DHInfra.at offers
Distinct infrastructure parts, contributed by partners across Austria. Open each one for details and contacts.
GPU & LLM Cluster
Train and fine-tune models, run interactive GPU notebooks, and call a hosted OpenAI-compatible LLM and embedding API — built for Digital Humanities, not generic HPC.
Read moreDigitization
Multispectral imaging (XpeCAM, MISHA), a scanning robot for fragile loose-leaf collections, and borrowable inference GPUs — available on request.
Read moreSaaS / IaaS & storage
Database-as-a-service, dev/test VMs and CI runners, plus short and mid-term project storage — a development-first complement to your own institution.
Read moreOpen Source Software
Three DH tools improved through DHInfra investment — QLever, liiive.now and Zellij. Usable on our cluster and elsewhere.
Read moreThe DHInfra.at consortium
9
Institutions
4
Working groups
50+
Members
100+
Associates
Use cases & news
View all posts »Real projects on our infrastructure, announcements, and reports from across the consortium.
Three and a half years of DHInfra.at: results, and the handover to regular operations
On 8 June 2026 DHInfra.at presented the results of three and a half years of building shared DH infrastructure across Austria. As the project closes at the end of June, a CLARIAH-AT follow-up takes over the rollout into live operations.
Reading the unreadable: multispectral imaging of medieval Serbian charters
DHInfra coordinated a loan of TU Wien's portable MISHA system to recover faded Old Serbian text on damaged charters in Cetinje — here is how the process works.
GPU4U in Action: Large-Scale Video Retrieval with Vision Language Models
How Ferdinand Köppen used a H200 GPU to describe nearly a million video keyframes for his master thesis.
Lacking resources or skills for your DH project?
We are open to project collaborations and happy to help in joint projects. Contact us and we will try to find the best solution for your idea.