Virtual Scientists leverages GPT-based language models to create specialized AI Agents that replicate expert scientists across various fields. Each virtual researcher is configured with tailored prompt engineering to provide accurate, context-aware answers, propose experimental protocols, interpret scientific data, and generate insights. Users select a scientific persona, input their questions or project details, and receive detailed, discipline-specific guidance supported by references and reasoning for educational or research purposes. The platform is hosted on GitHub Pages and is fully open-source. The codebase supports easy customization and extension of new scientific personas by modifying JSON configuration files. Ideal for collaborative research, teaching demonstrations, or personal study, Virtual Scientists bridges the gap between AI language models and practical scientific problem-solving by offering a dynamic, interactive environment for exploring complex topics with expert-like guidance.

Virtual Scientists V2 serves as a modular AI agent framework tailored for scientific research. It defines multiple virtual scientists—Chemist, Physicist, Biologist, and Data Scientist—each equipped with domain-specific knowledge and tool integrations. These agents utilize LangChain to orchestrate API calls to sources like Semantic Scholar, ArXiv, and web search, enabling automated literature retrieval, contextual analysis, and data extraction. Users script tasks by specifying research objectives; agents autonomously gather papers, summarize methodologies and results, propose experimental protocols, generate hypotheses, and produce structured reports. The framework supports plugins for custom tools and workflows, promoting extensibility. By automating repetitive research tasks, Virtual Scientists V2 accelerates insight generation and reduces manual effort across multidisciplinary projects.