SciDaSynth is an interactive system developed to improve the extraction of structured data from scientific literature by utilizing large language models. This technology addresses the challenge of handling diverse and inconsistent information commonly found in scientific texts, which can hinder researchers' ability to access critical data efficiently. By leveraging advanced language models, SciDaSynth facilitates more accurate and streamlined data extraction, supporting evidence-based decision-making processes. The system is designed to be user-friendly, enabling researchers to interactively refine and extract relevant data. According to claims associated with the system, SciDaSynth demonstrates positive effectiveness in achieving its goals. This development aligns with ongoing efforts in the field of natural language processing to harness large language models for practical applications in scientific research. The innovation holds promise for enhancing data accessibility and usability in scientific domains, as documented in recent literature on arXiv.