SteelScientist is a reliable AI expert based on large language models that represents an end-to-end pipeline from materials text to properties, enabling quantitative predictions of mechanical properties including yield strength (YS), ultimate tensile strength (UTS), and total elongation (EL) with high accuracy, as well as the exploration of new steels. The pipeline comprises a materials language encoder named SteelBERT, which is pretrained on a comprehensive corpus comprising 4.2 million abstracts related to materials science and 55,000 full texts from the steel literature, along with a multimodal deep learning framework that maps the composition and text sequence of complex fabrication processes to mechanical properties. This model achieves R2 scores of about 80% on test data for mechanical properties prediction.