Android-Based English Vocabulary Learning Media using Speech Recognition and Augmented Reality

English is one of the international languages that must be learned as early as possible, and vocabulary is the main obstacle in learning English. Children tend to like things related to games, so innovation in interactive English learning media is needed to attract children. The Tupai Pintar application offers solutions using three features: object detection, speech recognition, and speech assessment. This application has a sequential working principle from object detection to pronounce assessment. Children can use the application by exploring objects around them to scan and train English vocabulary pronunciation accompanied by the character Tutu Si Tupai Pintar. The Tupai Pintar App has two testing schemes. First, the Tupai Pintar Application has been tested on fifteen objects and has an object detection accuracy of 100%. The Tupai Pintar application has also been tested on elementary school children in grades 3 - 6 and has an average increase in pre-test and post-test scores of two points. This application can add more object datasets and increase interaction between characters and users for further development.