Wisp Science is an open-source, local-first desktop and CLI workbench designed for AI-assisted scientific research. It integrates papers, local data, Python/R scripts, research databases, and agent workflows into unified, traceable projects with persistent session history and reproducible outputs. Key capabilities include multiple agent delegation modes (Manual, Assisted, Automatic), built-in research skills, MCP database connectors, and local SQLite-based artifact tracking. It is built for researchers who need reproducible, self-contained scientific computing environments without relying on c
In the AI space, Wisp Science takes a focused approach. It focuses on managing scattered research artifacts like papers, local datasets, code, and agent workflows without reproducible, traceable project records. It is built as an open-source project for researchers. The project is open source (Apache-2.0). Wisp Science is available on the web, macOS, Windows, Linux, and the command line, and it can be self-hosted.
Behind Wisp Science is xuzhougeng, based in China, and it first shipped in 2026. Development happens publicly on GitHub with 194 stars and 718 commits in the last 90 days. Among its 6 catalogued features are Reproducible Workflows, Agent Delegation Modes, and Persistent Python/R Environment. The interface is available in English and Chinese. It exposes integrations via an MCP server. Wisp Science is currently in beta.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do