Abstract Objective To evaluate the efficacy and safety of Janus kinase inhibitors (JAKis) in JDM. Methods We conducted a single-centre retrospective study of patients with JDM treated by JAKi with a follow-up of at least 6 months. Proportion of clinically inactive disease (CID) within 6 months of JAKi initiation was evaluated using PRINTO criteria and skin Disease Activity Score. Serum IFN-α concentration was measured by Simoa assay. Results Nine refractory and one new-onset patients with JDM treated with ruxolitinib (n = 7) or baricitinib (n = 3) were included. The main indications for treatment were refractory muscle involvement (n = 8) and ulcerative skin disease (n = 2). CID was achieved in 5/10 patients (two/two anti-MDA5, three/four anti-NXP2, zero/three anti-TIF1γ-positive patients) within 6 months of JAKi introduction. All responders could withdraw plasmatic exchange, immunoadsorption and other immunosuppressive drugs. The mean daily steroid dose decreased from 1.1 mg/kg (range 0.35–2 mg/kg/d) to 0.1 (range, 0–0.3, P = 0.008) in patients achieving CID, and was stopped in two. Serum IFN-α concentrations were elevated in all patients at the time of treatment initiation and normalized in both responder and non-responder. A muscle biopsy repeated in one patient 26 months after the initiation of JAKi, showed a complete restoration of muscle endomysial microvascular bed. Herpes zoster and skin abscesses developed in three and two patients, respectively. Conclusion JAKis resulted in a CID in a subset of new-onset or refractory patients with JDM and may dramatically reverse severe muscle vasculopathy. Overall tolerance was good except for a high rate of herpes zoster infection.