$eosc-aissistant --status=active --grant=101292566 --duration=36mo --partners=12
agent.observe() →
⬡materials_science
∿energy_systems
Ψbiodiversity
Horizon Europe · HORIZON-INFRA-2025-01-EOSC-05

Agentic GenAI Open Research Assistant

A trustworthy, FAIR-by-design assistant to help scientists discover data, generate hypotheses, orchestrate workflows, and produce reproducible research outputs inside the EOSC ecosystem.

project.duration = 36 months

The project translates agentic AI into practical EOSC services for scientists, with human oversight at every critical decision point.

objectives.md

Core objectives

  • 01 Improve FAIRness, data quality, provenance, and bias detection across research domains.
  • 02 Develop trustworthy GenAI models for scientific data workflows and reasoning tasks.
  • 03 Build a modular agentic assistant for human–AI collaboration across the research lifecycle.
  • 04 Validate the methods through biodiversity, energy systems, and materials science use cases.
About the project

Trustworthy AI, built for open science

EOSC AIssistant is a European research project developing an agentic, AI-powered open research assistant fully integrated into the European Open Science Cloud (EOSC).

Coordinated by TIB – Leibniz Information Centre for Science and Technology with 12 partners across 7 countries, the project unites Generative AI, FAIR Digital Objects with RO-Crates, and cognitive Knowledge Graph technologies to support researchers throughout the entire research lifecycle — from hypothesis generation, literature synthesis, data discovery and FAIRification to provenance tracking, experimental design, and data management.

Validated through real-world use cases in biodiversity, energy system simulation, and materials science, EOSC AIssistant is built on a human-in-the-loop model that keeps researchers firmly in control of scientific judgement, ethics, and decision-making. All outputs are designed to be transparent, provenance-rich, and EOSC-compliant, fully aligned with Open Science principles.

12
partners
07
countries
36
months
vision.json
{
  "vision": "A Collaborative European Research System in which a Web of FAIR data and AI services enable trustworthy, reproducible discoveries.",
  "mission": "Develop, deploy, and pilot the EOSC AIssistant — an agentic, AI-powered research assistant integrated into the EOSC Federation via an EOSC Node.",
  "capabilities": [
    "find_fair_data",
    "process_knowledge",
    "manage_lifecycle"
  ],
  "oversight": "human_in_the_loop"
}
Ψ

Biodiversity

Biodiversity research produces vast, fragmented data — species observations, taxonomic records, ecological surveys — scattered across incompatible repositories and formats. EOSC AIssistant helps researchers discover, connect, and FAIRify this data, using agentic workflows to support hypothesis generation around species distribution, ecosystem change, and conservation questions. Naturalis Biodiversity Center anchors this use case, testing the assistant's ability to reason over biological knowledge graphs while keeping researchers in control of scientific judgement.

elements: species_records · taxonomic_graphs · ecological_surveys
⬡

Materials Science

Materials science discovery depends on connecting experimental results, simulations, and literature that too often live in disconnected silos. EOSC AIssistant supports researchers in navigating this landscape, orchestrating agentic workflows that retrieve relevant FAIR datasets, track provenance, and assist in reasoning across materials data. This use case validates the assistant's capacity to accelerate hypothesis generation and experimental design while keeping every step transparent and reproducible.

elements: crystal_structures · experimental_data · simulation_runs
∿

Energy Systems

Energy system simulation means modelling complex, interconnected infrastructure — generation, storage, distribution — under increasingly dynamic and uncertain conditions. EOSC AIssistant assists researchers in managing and interpreting simulation data, coordinating agentic workflows that support scenario exploration and reproducible analysis. This use case tests the assistant's ability to handle high-throughput, multi-domain simulation data with human oversight built into every critical decision.

elements: grid_topology · scenario_models · simulation_runs
Consortium

12 partners, 7 countries, one mission

A multidisciplinary consortium spanning research infrastructures, universities, and AI research organisations across Europe.

News & Events

Latest updates

Follow the project's milestones, meetings, and announcements.

1-2 October 2026 · hannover

EOSC AIssistant kicks off in Hannover to build AI research assistants that scientists can trust

The EOSC AIssistant consortium, twelve partners from seven countries, met in Hannover on 1–2 October 2026 to launch a three-year Horizon Europe project that will build a trustworthy AI research assistant for the European Open Science Cloud (EOSC).

read_more →
1 Sept 2026 · launch

EOSC AIssistant Project Officially Launches

TIB Leibniz Information Centre for Science and Technology is pleased to announce the official start of the EOSC AIssistant project on 1 September 2026. Coordinated by TIB and bringing together 12 partners from 7 European countries, the 36-month project will develop, validate, and deploy an agentic, AI-powered open research assistant fully integrated into the European Open Science Cloud (EOSC). Combining Generative AI, FAIR Digital Objects, and Knowledge Graph technologies, EOSC AIssistant will support researchers across the entire research lifecycle, from hypothesis generation to publication, with real-world validation in biodiversity, energy system simulation, and materials science. The project is funded under the Horizon Europe call HORIZON-INFRA-2025-01-EOSC-05. Further updates, including the kick-off meeting and project milestones, will follow on this website.

Contact

Get in touch

For questions about the project, partnerships, or media enquiries.

coordinator
TIB – Leibniz Information Centre for Science and Technology
address
Welfengarten 1 B, 30167 Hannover, Germany
principal investigator
Prof. Dr. Sahar Vahdati
administrative questions
Gino Erkeling
↑