Biomolecular temperature sensors can be used for efficient control of large-volume bioreactors, for spatiotemporal control and imaging of gene expression, as well as to engineer robustness to temperature in biomolecular circuit design. While RNA-based sensors, called 'thermometers', have been investigated in natural and synthetic contexts, an important challenge is to design different responses to temperature, differing in sensitivities and thresholds. We address this issue using experimental measurements in cells and in cell-free biomolecular 'breadboards' in combination with computations of RNA thermodynamics. We designed a library of RNA thermometers, finding, computationally, that it could contain a multiplicity of responses to temperature. We constructed this library and found a wide range of responses to temperature, ranging from 3.5-fold to over 10-fold in the temperature range 29°C - 37°C. These were largely linear responses with over 10-fold difference in slopes. We correlated the measured responses with computational expectations, finding that while there was no strong correlation in the individual values, the overall trends were similar. These results present a toolbox of RNA-based circuit elements with varying temperature sensitivities.