wisp-science

Why we included this project

Wisp Science is a desktop workbench for people who want an AI assistant to run their analysis rather than just talk about it. Each project keeps its own persistent Python and R environment, so variables survive between cells and even across restarts, and you can point that environment at local hardware, WSL, SSH, or a GPU box. The bundled MCP servers reach roughly eighty bioinformatics databases such as PubMed and GEO, loading them on demand so they don't bloat every request, which makes the workbench useful for computational biology. It also loads reusable Agent Skills, writes files only inside a sandboxed project root behind approval gates, and can drive external coding agents over ACP. Since data, conversations, and credentials stay on your own machines.