xuzhougeng/wisp-science

Open-source, local-first desktop AI research workbench for scientific computing with Python/R, MCP bioinformatics tools, SSH/WSL/GPU runtimes, and OpenAI/Anthropic models.

What it solves

Scientists often juggle literature search, coding in Python or R, querying multiple databases, and managing messy project records across different tools. This project brings those tasks into one desktop app that keeps all data, conversations, and credentials on your own machines instead of a cloud service.

How it works

You install the app on Windows, macOS, or Linux. You can plug in your own AI models (OpenAI-compatible, Anthropic, or Codex/Claude Code over ACP). The built-in agent reads and writes project files, runs shell commands, and loads reusable skills without flooding the prompt. You get persistent Python and R kernels per conversation, and you can register local or remote compute hosts to run long jobs. It also bundles MCP servers to search PubMed, GEO, and about 80 other scientific databases. Everything from figures, runs, decisions, and drafts is stored in one project, and the full history is restored on restart. You can manually sync or transfer projects with encryption, and no background sync happens.

Who it’s for

Researchers and scientists who want a private, local-first workbench for literature review, data analysis, and reproducible research, while having control over their models and data.

Highlights

  • Bring your own AI model and keep full control.
  • Run Python and R in isolated per-conversation kernels.
  • Query ~80 scientific databases via built-in MCP servers.
  • Approval gates for agent actions, with optional full permission.
  • Offline previews for notebooks, PDFs, Office files, and images.
  • Isolated explorations to try ideas without touching the mainline.
  • Publication Workspace that freezes manuscript revisions into verifiable Evidence Capsules.
  • Encrypted manual sync and project transfer—no background sync.

Related

  • Project
    aipoch/open-science

    An open-source, local-first AI research workbench that enables reproducible science through scientific AI agents, integrated Python/R execution, and traceable artifact provenance.

  • Project
    synthetic-sciences/openscience

    OpenScience is an open‑source AI research workbench that lets users describe scientific tasks in plain language. The built‑in agent plans, searches literature, writes and runs code (Python/R/shell), executes experiments on local or remote compute, and produces a reproducible report, with every step logged. Available as a desktop app, web workspace, or CLI, it supports custom LLMs, a large library of domain‑specific skills, and extensible plugins.

  • Project
    ai4s-research/open-science

    Open Science Desktop is a cross‑platform, local‑first AI workbench that lets researchers run the entire scientific workflow—from idea to paper—using autonomous agents. It offers full provenance, model‑agnostic runtime, split‑pane UI, remote/mobile access, and a powerful CLI (`osd`). Installers are provided for macOS, Windows and Linux, and the project is MIT‑licensed.

  • Project
    handsome-rich/Awesome-Auto-Research-Tools

    A curated directory of open-source projects that automate the scientific research process, including literature review, experiment execution, and paper writing.

  • Project
    swjybky/deepwrite

    A local-first AI writing workbench for creators that integrates specialized agents and a diff-based review system to collaboratively draft and refine manuscripts.