Predicting nanobody-protein interactions directly from amino acid sequences without the need for structural modeling. Sequence based predictions have potential advantages over structure based ones in terms of scalability, compute speed, and cost.