Open Science Desktop is a local-first, model-agnostic AI research workbench for macOS, Windows, and Linux.
The tool addresses the need for reproducible, auditable AI-assisted scientific research by integrating agents, notebooks, files, figures, reports, runs, and review into a single desktop workflow. It operates as an open-source alternative to proprietary AI-for-science platforms, built on Tauri for cross-platform desktop delivery, the Model Context Protocol for extensibility, and agent skills for automating research tasks. The local-first architecture means research data and computations remain on the user's machine rather than being sent to external services.
Researchers conducting computational science, data analysis, or experimental work who want to maintain control over their data and avoid vendor lock-in should consider this tool. It suits projects where reproducibility and auditability matter—where the ability to review and verify every step of an AI-assisted analysis is important. The tool is model-agnostic, meaning it can work with different AI models rather than being tied to a single provider, which distinguishes it from proprietary alternatives that depend on specific commercial services.
The project maintains active development across its core functionality, with ongoing work to expand agent capabilities and improve the research workflow integration. The tool supports seven interface languages, indicating sustained effort toward accessibility for international research communities. The codebase is built with TypeScript, Tauri 2, and React, reflecting a modern approach to desktop application development.