Plastics have invaded human lives to an extent where India alone generated 9.3 million tons of plastic waste annually. The molecular complexity of plastics, coupled with their mass presence and the fallacies in the disposal process has errored multiple dimensions of our ecosystems. Currently, the most prevalent Municipal Solid Waste (MSW) disposal technique is waste dumping in landfills foreboding a threat to human health and diluting the global sustainability ventures. Further analysis of the consumption pattern suggests the dominance of Polyethylene Terephthalate (PETE) and High-Density Polyethylene (HDPE) as the most extensively recycled grades with exclusive Polypropylene (PP), Low-Density Polyethylene (LDPE), and Polyvinyl Chloride (PVC) recycling. Grades 6 and 7 are the least recycled grade and consist of objects that particularly lie in the non-recyclable category. This non-uniformity in the recyclability of plastic grades stands as the major hindrance to recycling technology. India solely relies on rag-pickers and scrap dealers for the supply of segregated plastics which is biased and flawed due to extensive human dependence. This paper discusses the implementation of an open-source, computer vision-based plastic segregation using an XYZ Cartesian robot with an adaptive gripper for accuracy in grade-wise segregation of plastic waste. A YOLOv8 machine learning model, trained on a customized dataset, conducts unsupervised learning for the recognition of the different plastic grades. The novelty of this research lies in building an economical and accurate solution which highlights the versatile possibilities of an industry-level scale-up further strengthening the implementation of a sustainable and circular economy.