This tool performs RAA on scientific texts to extract mentions of research artifacts (RAs) (e.g. datasets, software) along with their metadata and then deduplicates these mentions to find the unique RAs that were referenced, (re)used or created in the text. It is based on finetuned LLMs and depends on a list of keywords/keyphrases/gazetteers to detect candidate mentions to RAs in the texts (source: default list of keywords for Computer Science and a default list of gazetteers from PapersWithCode). For more information: Github: https://github.com/iNoBo/scinobo-raa HuggingFace Space: https://huggingface.co/spaces/iNoBo/scinobo-research-artifact-analysis Dockerhub: https://hub.docker.com/repository/docker/intelligencenoborders/scinobo-raa