Transition-metal complexes are used as photosensitizers1, in light-emitting diodes, for biosensing and in photocatalysis2. A key feature in these applications is excitation from the ground state to a charge-transfer state3, 4; the long charge-transfer-state lifetimes typical for complexes of ruthenium5 and other precious metals are often essential to ensure high performance. There is much interest in replacing these scarce elements with Earth-abundant metals, with iron6 and copper7 being particularly attractive owing to their low cost and non-toxicity. But despite the exploration of innovative molecular designs6, 8, 9, 10, it remains a formidable scientific challenge11 to access Earth-abundant transition-metal complexes with long-lived charge-transfer excited states. No known iron complexes are considered12 photoluminescent at room temperature, and their rapid excited-state deactivation precludes their use as photosensitizers13, 14, 15. Here we present the iron complex [Fe(btz)3]3+ (where btz is 3,3′-dimethyl-1,1′-bis(p-tolyl)-4,4′-bis(1,2,3-triazol-5-ylidene)), and show that the superior σ-donor and π-acceptor electron properties of the ligand stabilize the excited state sufficiently to realize a long charge-transfer lifetime of 100 picoseconds (ps) and room-temperature photoluminescence. This species is a low-spin Fe(iii) d5 complex, and emission occurs from a long-lived doublet ligand-to-metal charge-transfer (2LMCT) state that is rarely seen for transition-metal complexes4, 16, 17. The absence of intersystem crossing, which often gives rise to large excited-state energy losses in transition-metal complexes, enables the observation of spin-allowed emission directly to the ground state and could be exploited as an increased driving force in photochemical reactions on surfaces. These findings suggest that appropriate design strategies can deliver new iron-based materials for use as light emitters and photosensitizers