RAISE-CONNECT ( HORIZON-INFRA-2025-01-EOSC-02-101292695) develops AI-powered, domain-agnostic connectors that enable digital research objects—including data, software, experiments and results—to become FAIR, interoperable and reusable across repositories and data spaces. Building on existing EOSC-RAISE technologies, the project will introduce automated FAIRification services, a common repository connector framework and natural-language interfaces that allow researchers to discover, access and process digital objects with minimal technical effort. The solutions will be validated across more than 12 real-world repositories from diverse scientific domains, supporting reproducibility and making Open Science practices easier to integrate into everyday research workflows.
Despite significant progress in open data, code sharing, and implementation of FAIR principles, there is no automated, universal mechanism capable to “FAIR-ify” DOs across large cross-domain repositories and data spaces. RAISE-CONNECT addresses this challenge by providing AI-powered, domain-agnostic connectors that can “FAIR-ify” DOs (i.e. data, code, experiments, experimental results and FDOs) and make them interoperable both within and across repositories. Leveraging existing EOSC-RAISE tools, it offers researchers and other stakeholders an entry point to access, share, and process FDOs with no effort. RAISE-CONNECT establishes an architectural framework for seamless creation, transformation, and integration of FDOs. At its core, it introduces an automated “FAIR-ification” service to continuously enhance DOs by searching across repositories for missing FAIR elements. This ensures that every DO can be transformed into an FDO that is discoverable, interoperable, and reusable. To maximise usability and uptake, RAISE-CONNECT lets users interact with DOs through natural language prompts, leveraging LLMs: users can “ask” for datasets, scripts, compare results, all without technical overhead. A model of Common Repository Connector acts as ”template” abstraction layer to bridge heterogeneous repositories, while different specific implementations of the it will be developed for well-known Data Spaces and DO repositories. The FAIR-ification Engine provides continuous and on-demand FAIR-ness assessment and transformation. Complex workflows can be initiated by natural language queries or graphical tools. By removing technical barriers and enforcing compliance, provenance, and reproducibility by design, it turns open science from a policy into an everyday practice for researchers, innovators, and data providers. RAISE-CONNECT will be validated on more than 12 real-life data, code, and experiment repositories, ensuring its applicability across diverse domains.
