The increasing volume of scientific research necessitates efficient and accurate methods for processing and disseminating knowledge across languages. This paper presents a pipeline for the comprehensive translation of scientific documents, addressing crucial gaps in existing open-source solutions. Our pipeline encompasses four key stages: PDF-to-image conversion for enhanced document layout analysis, robust text recognition, language-specific translation tailored to scientific terminology, and document reconstruction. Our pipeline establishes a way for one-click translation of scientific documents from other languages into English, addressing a critical need in the scientific community. The pipeline comprises four stages: PDF-to-image conversion, robust text recognition, language-specific translation, and document reconstruction. Text recognition is enhanced through image processing techniques, while translation is tailored to scientific terminology using domain-specific translation models. Experiments demonstrate significant improvement of translation accuracy compared to other open-source models for Russian reaching BLUE score 45 for Scientific documents . We demonstrate the efficacy of our approach through experiments, showing significant improvements in translation accuracy by significantly reducing errors in scientific terminology and by predicting optimized document structure.