Purpose – Blind people have limitations in accessing the outside world using their sense of sight. One of them is AR for the compound blind. This research aims to train AR to read and write braille using the left-code method. Method – The methods used are experimental methods with a Single Subject Research (SSR) research design with an A-B-A design. Data collection uses tests. The data obtained was analyzed using descriptive statistics and displayed as tables and graphs. Findings – Using the left method technique, it was discovered that AR subjects could read and write braille for packages 1 to 3. An increase in the average level in each condition showed this. Starting from the baseline condition A1, which is a score of 1.3, the intervention condition gets a score of 9.6 and the baseline condition A2 gets a score of 16. So AR can read syllables, words, and sentences in package 1: a,b,k,l, package 2:c,f,m,p, package 3: e,h,o,r. Research Implications – The implications of this research indicate that the left coding method can be used to train compound blind subjects to learn braille. Time limitations, limited learning facilities, and limited intellectual conditions of the subjects mark the limitations of this research. Researchers suggest that for further research, considering the characteristics of compound blind subjects who are easily suspicious and irritable and have below-average intelligence, learning to read and write braille must be done in a calm situation. The condition of a comfortable place without interference from other people's voices also needs to be considered.