Learning media with object detection and speech recognition features represent innovative approaches in facilitating the learning process. This research aims to increase student enthusiasm for learning and introduce new innovations for learning media in elementary schools, achieved through the implementation of the SREDO application. The research method employed is Research and Development (R&D) utilizing the 4D development model (Define, Design, Development & Disseminate). SREDO serves as a learning media application used for recognizing fruit names in English, accomplished either through student voice or direct scanning of fruit images. The media was developed using MIT App Inventor and Google Speech software. The research incorporates a series of evaluations to assess functionality, compatibility, and validity. Functionality testing is conducted through blackbox testing, yielding a 100% success rate, while compatibility testing with Firebase Test Lab on five different devices also achieves a 100% success rate. Expert evaluation by media specialists resulted in an 87% approval rating, indicating the validity and suitability of the media. Evaluation by a material expert, an English teacher, yielded a score of 89%, further affirming the media's appropriateness for learning objectives. Additionally, evaluation by 33 students using the System Usability Scale (SUS) resulted in an average score of 84.09, signifying a highly satisfactory and acceptable level of usability. Overall, the developed SREDO media application is proven to be highly suitable and effective for introducing fruit names in English within elementary schools.