Adolescent idiopathic scoliosis (AIS) is the most pervasive spinal disorder in the adolescence. Clinical screening and diagnosis require manual physical and radiographic examinations, which is either manual and subjective, or increases the radiation exposure to children to analyse the alignment of the spine. Thus, we developed and validated a radiation-free system and device utilising artificial intelligence (AI) and depth sensing (RGBD) technologies to analyse AIS. Our in-house developed device takes an RGBD image of the nude back as input and synthesises its spine alignment (Figure 1). The synthesised spine alignment contains sufficient anatomical information that can assist disease severity and curve type prediction. We tested the performance of our system prospectively at two local clinics on 230 patients. The synthesised spine alignment for AIS severity classification achieved a sensitivity and NPV of 0.95 and 0.97, and these performances were 0.89 and 0.86 for the curve type classification, with spine specialists’ manual assessment results from real radiographies as ground truth (GT). Our portable system and device have the potential to assist fast and accurate AIS screening without radiation.