New open-source software uses deep learning to automate the quantitative assessment of Fusarium root rot in cereal seedlings, overcoming the limitations of manual scoring.