In this work, we introduce NMRexp, an open and curated database comprising 3.3 million experimental NMR spectra across six nuclei (1H, 13C, 19F, 31P, 29Si, and 11B), extracted from nearly 200,000 Supporting Information documents published between 2010 and 2024. By combining large-scale automated extraction with strict rule-based parsing and chemical-consistency validation, NMRexp achieves > 99% accuracy in metadata extraction (chemical shifts, multiplicities, J-couplings, solvents) and 98% correctness in molecular-skeleton assignment; internal consistency tests yielded mean absolute errors of 0.026 ppm for 1H and 0.206 ppm for 13C chemical shifts.
