In this paper, we present a framework to improve robotics' perception capabilities by combining traditional computer vision techniques with Google Cloud Vision API. By integrating a vast ontology of over 800 object classes, adaptation across domains can be achieved, obviating the necessity for environment-specific technologies. Although recent developments in computer vision and sensor technologies have enhanced robotics perception, it is still difficult to achieve accurate object detection on hardware with limited resources. Our technology offloads laborious processes by utilizing cloud computing, guaranteeing real-time responsiveness even on limited hardware. Through the use of cloud-vision, the system achieves increased accuracy when recognizing specific objects in the surroundings.