FAIRenrich: Empowering FAIR and Scalable Annotation Workflows in Earth System and Biodiversity Sciences

FAIRenrich is an open-source software solution that facilitates FAIR and scalable annotation of large, multilingual textual datasets in Earth System Sciences (ESS) and Biodiversity research. Designed with user-driven workflows in mind, it combines automated and human-guided terminology annotation using advanced natural language processing (NLP) models, including spaCy, Ollama, and GPT4All, as well as the external TIB Terminology Service. Users have full control over the annotation process via a user-friendly web interface that supports both batch processing of CSV files and real-time annotation. They can flexibly choose terminology sources, review and validate annotation results, and generate detailed statistics. A built-in caching mechanism optimizes performance and reduces reliance on external services. Due to its modular and highly configurable architecture, FAIRenrich seamlessly integrates local and remote AI models. By enabling transparent, reproducible, and customizable annotation workflows, FAIRenrich supports the FAIR principles and enables researchers and data stewards to manage and semantically enrich complex multilingual datasets—fostering open science and interdisciplinary collaboration in ESS and Biodiversity domains.