Designing biosensors to detect novel molecules is important in biotechnology. Allosteric transcription factors (aTF), which are widely used as biosensors, have proven challenging to design because mutating ligand-binding residues often disrupt allostery. Moreover, existing screening methods lack the sensitivity to identify rare, functional variants amidst a large pool of non-functional designs. We introduce Sensor-seq, a platform for designing and screening aTF biosensors with non-native ligand specificity. Sensor-seq uses phylogeny-guided library design, implicitly capturing allostery, and employs an RNA barcoding system for sensitive, highly multiplexed screening. From a library of 17,737 variants of aTF TtgR, Sensor-seq identified novel biosensors for six non-native ligands with high dynamic range and diverse specificity profiles. The structure of a naltrexone-bound design showed shape complementary methionine-aromatic interactions driving ligand specificity. To demonstrate practical utility, we developed cell-free detection systems with our designed biosensors. Sensor-seq enables rapid, scalable design of new biosensors, overcoming constraints of natural biosensors.