Optical Character Recognition (OCR) plays an important role in document image processing. Recognition of characters in a smart way is gaining importance in the modern days, as huge piles of data is generated, and it needs to be processed and manipulated. OCR eliminates the need of retyping in case of erasure of an important digital file accidentally, it quickens digital searches, enables faster text editing, saves up the storage space with the use of efficient algorithm. Accessibility of OCR to the visually impaired is a great progress, Where users can scan books, magazines, incoming mails or other programs. OCR can also be coupled with a voice-over utility, or voice synthesizer, that reads out the text recognized optically by the software. OCR also has its own bottlenecks. It may not recognize complex text, mathematical variables, hand written scripts. High end OCR utilities are used in areas where required, but it comes with an additional computational cost. This paper aims at presenting an OCR utility which recognizes text characters, using a machine learning model.