SciStudio

Scientific workflow bench with your AI partner.

English | 简体中文

What is SciStudio?

SciStudio is an interactive workflow orchestration system for multimodal scientific data analysis. You and your AI partner works together on a same canvas interactively. It starts from spatial multi-omics and is progressively expanding to other data modalities. SciStudio lets researchers connect scientific software, AI agents, scripts and multimodal data into one visual workflow.

• Typed data, one graph — blocks exchange typed data and flow through the same workflow, so every step stays connected end to end.
• View your data in your way - Customize the way you preview your data, customize your figure with plot cards.
• Bring your existing tools — run R or Python scripts and launch desktop apps like Fiji as ordinary blocks in the flow.
• AI-native — a built-in assistant (Claude Code or Codex) helps you build workflows, author new blocks, and inspect your data.
• Extensible — add your own blocks, data types, and plots, and share them as installable packages.

Install

For users

Download the latest SciStudio desktop app from the Releases page — a macOS .dmg or a Windows installer. Open it and you're ready; Python and all dependencies are bundled, so there is nothing else to set up.

Then follow the Quickstart.

For servers

pip install scistudio

The wheel bundles the web frontend. Every desktop update is also a PyPI release of the same build: 0.3.4-alpha-build0029 is scistudio==0.3.4a29. Once a stable release exists, add --pre to get the newest alpha.

For developers (run from source)

git clone https://github.com/jiazhenz026/SciStudio.git
cd SciStudio

# Python backend (use a conda env or a virtualenv)
python -m pip install ".[dev]"

# Frontend dependencies
npm --prefix frontend install

# Run the desktop app against your source:
# Vite HMR for the frontend + the SciStudio backend + Electron.
npm --prefix desktop run dev

Frontend edits hot-reload; restart the command to pick up backend changes. See desktop/README.md for packaging the app (.dmg / Windows installer).

Documentation

Full documentation lives at jiazhenz026.github.io/SciStudio:

• User Guide — building and running workflows, previewing data, history and branches, the AI assistant, and writing your own blocks, types, and plots.
• Quickstart — from a fresh install to your first running workflow.
• API Reference — the public API you can rely on, with signatures and stability tiers.
• Package Development — building a distributable SciStudio package (blocks, types, previewers).
• Architecture — how SciStudio is built and why.

The User Guide and API Reference are the same docs SciStudio provisions into each project, so what you read online matches what ships with the app.

Contributing

Contributions are welcome — bug reports, feature ideas, docs, and code. Start by reading CONTRIBUTING.md, and see AGENTS.md for the full development workflow (branch, issue, gate, tests, docs, review).

To build and ship your own blocks (rather than change the core), follow the Package Development guide.

Community

Questions, feedback, and bug reports are very welcome while SciStudio is in alpha:

• Discord
• GitHub Issues

Status

SciStudio is in alpha and under active development. Interfaces and APIs may change between releases; the API Reference marks the stability tier of each public symbol.

License

SciStudio is released under the Apache License 2.0. See LICENSE for the full text.