Erasmus+ KA220-HED · 2026 – 2029
RAG4Edu empowers educators and students to move beyond generic AI tools and build their own pedagogically grounded RAG applications — with a competence-based curriculum, an open training programme, accessible tools and a European Community of Practice.
About the project
Large language models are entering teaching and learning fast, but their use is often ad hoc, prompt-based and weakly anchored in educational knowledge. That raises real concerns about reliability, accuracy, bias, transparency and data protection.
Retrieval-Augmented Generation (RAG) combines generative AI with curated, domain-specific knowledge sources — lecture notes, slides, scientific publications, exams, lab materials — so that AI support in a course is more controllable, explainable and context-aware. Educators across Europe, however, currently lack the training, pedagogical frameworks and institutional support needed to design and run RAG-based solutions.
RAG4Edu closes that gap. It develops a competence-based RAG curriculum and a modular training programme with micro-learning modules, group-based learning activities, accessible tools, templates and worked examples that let educators create their own RAG applications without advanced AI skills. Cross-country pilots validate the programme, and a European Community of Practice carries it beyond the project.
The project promotes European, values-aligned AI — EU-based LLMs and EuroHPC infrastructure, open licences, GDPR-compliant design and alignment with the EU AI Act — contributing to a digitally competent and sovereign European higher-education sector.
How RAG works
RAG combines generative AI with the knowledge sources an educator chooses, so AI support in a course is controllable, explainable and context-aware.
A learner’s question searches the course knowledge the educator has curated — lecture notes, slides, papers, exams — and returns the most relevant passages.
The retrieved passages and the learner’s profile are added to the prompt, so institutional knowledge shapes a relevant, context-aware response.
A European LLM writes the answer from that context and shows the sources it used, so learners can check every claim.
Educators design and validate each step in the RAG4Edu Toolbox — with GDPR and AI Act checks, and without advanced AI skills.
Objectives
RAG4Edu strengthens the capacity of European higher-education institutions to design, implement and evaluate RAG for teaching and learning in a pedagogically grounded, responsible and sustainable way.
European AIPromote European AI technologies — EU-based LLMs such as EuroLLM, Mistral and Aleph Alpha, and EuroHPC infrastructure — with transparency, explainability, accountability and data sovereignty, in compliance with the AI Act.
PedagogyAnalyse the AI-in-education landscape, synthesise the state of RAG, and design the RAG4Edu Curriculum with competence levels, learning outcomes, instructional principles and micro-credentials.
ToolsBuild the RAG4Edu pipeline and framework, learner-aware augmentation, AI-supported group formation, and an authoring toolbox for multilingual educational chatbots on European LLMs.
TrainingDesign, pilot, evaluate and refine the RAG4Edu Training Programme in coordinated pilots across partner countries, generating evidence on effectiveness, usability and institutional feasibility.
CommunityRun the communication platform and a European Community of Practice, publish everything open access, and secure adoption through a sustainability and exploitation strategy, multiplier events and a final European workshop.
News and events
Updates from the partnership, upcoming webinars, co-design sessions and multiplier events.
One RAG4Edu account now gives access to the project website, its members' area and the RAG4Edu learning platform.
Read more →
NiceTransnational project meeting hosted by Université Côte d'Azur.
Read more →
LisbonTransnational project meeting hosted by INESC-ID.
Read more →
BarcelonaTransnational project meeting hosted by IIIA-CSIC.
Read more →Partnership
Higher-education institutions with strong teacher-education and digital-pedagogy profiles, together with two of Europe’s leading AI research centres.
Berlin, Germany · bht-berlin.de
One of Germany’s largest public universities of applied sciences, with around 11,600 students. Its Data Science +X Research Center includes an AI + Education strand, and the Competence Center for Digital Teaching supports the use of AI in teaching.
Prof. Dr. Ilona Buchem · Prof. Dr. Klaus-Peter Schoeneberg
Barcelona, Spain · iiia.csic.es
An internationally recognised AI institute of the Spanish National Research Council, working on machine learning, multi-agent systems and reasoning. It coordinates AIHUB-CSIC, a network of more than 500 AI researchers.
Prof. Carles Sierra · Prof. Margarida Romero · Dr. Mireia Vendrell
Nice, France · univ-cotedazur.eu
A public research university with about 35,000 students. Its LINE laboratory studies educational innovation, co-creativity and critical thinking, and its MSc SmartEdTech combines computer science, education and cognitive science.
Prof. Dr. Oksana Strutynska · Chengcheng Li
Oulu, Finland · oulu.fi
A Finnish research university founded in 1958. Its Learning and Educational Technology Research Lab (LET Lab) studies learning with multimodal data and human-AI collaboration, and the Hybrid Intelligence programme explores how humans and AI learn together.
Dr. Andy Nguyen · Prof. Sanna Järvelä · Dr. Luna Huynh · Azusa Nakata
Lisbon, Portugal · inesc-id.pt
A non-profit research institute in computer science and engineering, jointly owned by Instituto Superior Técnico and INESC. Its Group of Artificial Intelligence for People and Society (GAIPS) works on intelligent agents, social robotics and technology-enhanced learning.
Prof. Ana Paiva · Prof. Rui Prada · Prof. Joana Campos
Kyiv, Ukraine · udu.edu.ua
A long-established Ukrainian university with a strong tradition in teacher education and pedagogical research, and a growing focus on digital transformation, AI and educators’ digital competences.
Dr. Mariia Umryk · Daniil Shemet
Results
All core results are published under open licences in repositories such as Zenodo and GitHub, with persistent identifiers and FAIR metadata. Links will appear here as each result is released.
International policies, AI-literacy frameworks and research trends across the EU, China, the USA and partner countries; at least 50 sources reviewed.
In progressRAG architectures, tools and educational applications, a comparative RAG-tools matrix, and the core RAG competences that underpin the micro-credentials.
PlannedCompetence-based learning outcomes, instructional design principles, micro-learning modules, flexible pathways and micro-credentials as open badges.
PlannedCriteria, data sources, decision logic and transparency mechanisms for AI-supported group formation, with collaborative learning design templates.
PlannedConceptual and technical specification, a deployable operational prototype (ingestion, semantic retrieval, generation) and learner-aware augmentation mechanisms.
PlannedAuthoring tools to create and configure course-specific RAG chatbots and agents, multilingual case-study applications, templates and documentation.
PlannedSpecification and integration of EU-based LLMs and AI services, embeddings, token management and compliance features for the pilots.
PlannedOpen online programme combining expert-level RAG content, flexible pathways, OERs, competence-based assessment and micro-credentials.
PlannedParticipant selection, implementation guidelines, evaluation framework and data plan applied consistently across partner institutions.
PlannedEvidence from all pilots, lessons learned, and action points for improvements, translations and national adaptations.
PlannedThis website and the linked repositories — the project's open digital hub for outputs, events and dissemination material.
In progressA European Community of Practice with recurring thematic webinars, participatory co-design sessions and institutional roundtables.
PlannedAdaptation guidelines, licensing clarity and transfer models so other institutions can implement RAG4Edu results independently.
PlannedAt least three concrete post-project pathways: institutional teacher-training programmes, digital strategies and external networks.
PlannedA national multiplier event in every partner country and a final workshop as a pre-conference event at a leading AI-in-education conference.
PlannedCommunity of Practice
The RAG4Edu Community of Practice brings together educators, teaching and learning centres, innovation hubs and researchers across Europe — with webinars, co-design sessions and institutional roundtables from late 2026. Results are also shared through the European Digital Education Hub, EPALE, OpenAIRE and Zenodo.
One free RAG4Edu account gives you the members’ area of this website and the courses on the RAG4Edu learning platform.
