<!-- TITLE: LLMduckt at the LLT Breakfast Special: 20 years of Educational Technology -->
<!-- DESCRIPTION: The Educational Technology unit celebrates its 20th anniversary with a breakfast for teaching staff — and FLAAIT presents LLMduckt there. -->
<!-- LANG: en -->
<!-- URL: https://flaait.tugraz.at/en/news/2026/llmduckt-at-the-llt-breakfast-20-years/ -->
<!-- URL_DE: https://flaait.tugraz.at/news/2026/llmduckt-beim-llt-fruehstueck-20-jahre/ -->
<!-- URL_EN: https://flaait.tugraz.at/en/news/2026/llmduckt-at-the-llt-breakfast-20-years/ -->
<!-- TYPE: news -->
<!-- DATE: 2026-11-10 -->
<!-- DATE_PRECISION: day -->
<!-- PILLARS: ai -->
<!-- PILLAR_TITLES: AI for Teaching -->
<!-- SERVICES: llmduckt -->

# LLMduckt at the LLT Breakfast Special: 20 years of Educational Technology

*10 November 2026*

*The Educational Technology unit celebrates its 20th anniversary with a breakfast for teaching staff — and FLAAIT presents LLMduckt there.*

![A white box with the LLMduckt logo — a bridge over waves with a yellow duck — on a table; on top sits a rubber duck with a mortarboard and a diploma. Stickers on the box read “Don’t feed the model your data – stay on the LLMduckt”. In front lie yellow and blue TELucation notebooks reading “Lehre neu schreiben – mit TELucation.” (in German).](https://flaait.tugraz.at/_app/immutable/assets/llmduckt-20.BF5rAbG1.webp)

The Educational Technology unit at TU Graz turns 20 and is celebrating with an LLT Breakfast Special on 10 November 2026. In the short input, we present [LLMduckt](https://flaait.tugraz.at/en/services/llmduckt/): the gateway through which teaching staff give their students their own API keys for AI language models in the TeachCenter course. Afterwards there is time to talk with colleagues and the LLT team — over a shared breakfast.

- Tuesday, 10 November 2026, from 8 a.m.
- Lecture hall HS II, Rechbauerstraße 12
- Registration as usual by a short e-mail to [telucation@tugraz.at](mailto:telucation@tugraz.at)

![Architecture diagram of LLMduckt. In the centre, the gateway with four components: LTI tool (course, role and membership from the LMS, no separate accounts), course rules and quotas (model templates, per-person quotas, usage logged without prompt content), API keys per person and course, and one interface for many backends. On the left the Moodle LMS, teaching staff who choose model templates and limits, and students who send requests from their own code. On the right, in-house models on university GPU servers — request and response stay within the university boundary. Outside it, optional cloud models from contracted providers; only on this path does data leave the university.](https://flaait.tugraz.at/_app/immutable/assets/llmduckt-architektur.xN0AO_Wt.svg)

*How LLMduckt is built: the TeachCenter supplies course and role, the gateway manages keys and quotas and forwards requests to in-house models or — optionally — to the cloud.*

## Further links

- [Guide: Access to AI APIs with LLMduckt (in German) · telucation.tugraz.at](https://telucation.tugraz.at/anleitungen/zugriff-auf-ai-apis-mit-llmduckt/)
- [Invitation in the TeachCenter forum (in German) · tc.tugraz.at](https://tc.tugraz.at/main/mod/forum/discuss.php?d=120656)

## Related services

- [LLMduckt](https://flaait.tugraz.at/en/services/llmduckt/)