xuzhougeng/wisp-science
Wisp Science is pitched as a local-first AI workbench for scientific research, with models, data, credentials, code runs, and project history kept on the user’s machines.
The repo says Wisp can search literature, query PubMed, GEO, and about 80 other scientific databases, and run persistent Python and R kernels. Its agent can work with OpenAI-compatible or Anthropic models, or drive Codex and Claude Code over ACP, with approval gates unless the user enables full permission. It supports local, WSL, and SSH compute, live logs for longer runs, offline previews, project transfer, encrypted manual sync, and “Evidence Capsules” for manuscript revisions. Source: GitHub · LLM repos' note.
The repo says Wisp can search literature, query PubMed, GEO, and about 80 other scientific databases, and run persistent Python and R kernels. Its agent can work with OpenAI-compatible or Anthropic models, or drive Codex and Claude Code over ACP, with approval gates unless the user enables full permission. It supports local, WSL, and SSH compute, live logs for longer runs, offline previews, project transfer, encrypted manual sync, and “Evidence Capsules” for manuscript revisions. Source: GitHub · LLM repos' note.
score 4