Some of the most toxic substances known are the botulinum neurotoxins. The most well-known member of this toxin family is Botulinum neurotoxin A1, famously known as Botox. Botulinum neurotoxins are large proteins that target neurons and block the release of synaptic vesicles at nerve terminals, causing paralysis (1, 2). Yet despite their extreme toxicity, botulinum neurotoxins are used therapeutically and cosmetically. Expanding their potential beyond their natural substrates has been limited by the lack of tools to master the complex substrate specificity of the protease domains of these toxins. On page 803 of this issue, Blum et al. (3) describe a phage-assisted evolution approach to control the substrate specificities of the toxin’s protease domains. This may lead to new therapeutics, targeting diseases that previously have been inaccessible to treatment.