Authors - Ashwin Makwana, Hemit Rana, Adnan Vahora, Heer Patel, Shruti Rana, Nisarg Shah, Yagnik Poshiya Abstract - Although most of us are unaware of it, we all use Natural Language Processing -based software and applications in our daily lives to assist us in performing our responsibilities in the twenty-first century. The task of automatically transforming one natural language into another while keeping the same meaning of the input text and generating fluent text in the output language is known as machine translation (MT). While machine translation is one of the artificial intelligence's oldest subfields, the current shift toward large-scale empirical methodologies has resulted in considerable translation quality improvements. There are numerous features of MT that are difficult to master: the enormous number of languages, alphabets, and grammars; transforming a sequence (for example, a sentence) to a sequence is more difficult for a computer than working with numbers alone; there is no single correct answer in most of the cases. (e.g.: translating from a language without gender-dependent pronouns, he and she can be the same). Linguistic resources such as corpora, multilingual dictionaries, and morphological rules are needed for optimal Machine Translation. The attempt has been made to make semi-automatic dictionary generation tool using historical corpora. in this research, as dictionaries are a critical resource for machine translation. The task becomes easier when related scripts and languages are considered, and the developed project is aimed for scenarios in which there is a requirement bilingual semi-automatic dictionary generation tool.