Show HN: Open Science, open-source alternative to Claude Science (github.com)

An innovative open-source project has emerged as a direct alternative to Claude Science, known as Open Science. This AI workbench is designed for scientific research, integrating literature, code, figures, and reports into a cohesive and reproducible workflow. Unlike typical AI chat interfaces, it operates as a comprehensive environment that emphasizes traceable data and outputs, allowing researchers to conduct analyses with clear ties to the origin of every artifact produced. Open Science stands out for its model-agnostic approach, enabling users to utilize various AI models, including those from OpenAI and local options. It prioritizes data privacy by keeping computations and raw data local to the user’s machine while offering a transparent view of what data may be sent to external models. The platform supports a full research cycle—from planning and execution to artifact review—enhancing the integrity of scientific communication. With features such as a local Python kernel, built-in auditing tools, and a command palette for streamlined navigation, Open Science aims to reshape how researchers collaborate and document their findings, addressing the critical need for reproducibility in scientific work.