Zenodo hosts the archival version of this document; for convenient viewing, please visit andrewgyork.github.io/gfp_magnetofluorescence Abstract We've discovered a simple, nontoxic, biocompatible way to control the brightness of GFP-like fluorescent proteins via modest magnetic fields (~10 mT). Fluorescent proteins which seem magnetically inert (e.g. EGFP, mScarlet) become magnetoresponsive in the presence of an appropriate cofactor (e.g. EGFP-FlavinTag, or an mScarlet/FMN solution). This method works at room-temperature and body-temperature, in vitro, in E. coli and in cultured mammalian cells. The GFP-family magnetoresponse is weak (ΔF/F≈1%), but shows the hallmarks of evolvability. This suggests exciting technological possibilities, both short-term (e.g. lock-in detection, multiplexing) and long-term (e.g. optically-detected MRI, magnetogenetics). We've also discovered weak magnetoresponse from a member of the LOV-domain family. This suggests the possibility that magnetoresponse is a general feature of fluorescent proteins, and not unique to the cryptochrome/photolyase family. UPDATE: We've increased the magnetoresponse of AsLOV2 via directed evolution, yielding a protein (MagLOV) with a remarkably large ΔF/F≈75%. We consider MagLOV to be the most promising candidate ancestor for magnetogenetic tools.