Installation

Requirements

Python 3.10+
uv (strongly recommended) — install with pip install uv or curl -LsSf https://astral.sh/uv/install.sh | sh
pip 21+ also works, but expect 30–90 min for scitex[all] (see warning below)

Warning

pip install "scitex[all]" typically takes 30–90 minutes because pip’s serial resolver walks version histories of the large transitive dependency set (numpy/pandas/torch/jax/playwright/openalex-local/…). Use uv instead — it resolves the same set in parallel in 1–3 minutes. Every pip install line on this page also works as uv pip install and we recommend the uv form.

Core Install

The base package pulls in only lightweight dependencies and gives you access to session management, path utilities, string helpers, and the module discovery system.

uv pip install scitex          # recommended
pip install scitex              # also works, slower for [all]

Recommended: Install Everything

If you want the full experience (plotting, statistics, I/O, scholar, writer, and every other module):

uv pip install "scitex[all]"   # ~3 min
pip install "scitex[all]"       # ~30–90 min (resolver thrash)

Research Workflow

For a typical research project you need figures, statistics, and literature search but not audio, browser automation, or cloud tools:

pip install "scitex[plt,stats,scholar]"

Per-Module Extras

Install only what you need. Each extra maps to a self-contained capability.

Extra	Description
`plt`	Publication-ready figures via figrecipe (matplotlib, seaborn, Pillow)
`stats`	Hypothesis testing, effect sizes, power analysis (scitex-stats, scipy, statsmodels)
`io`	Unified I/O for 40+ formats (HDF5, Excel, YAML, PDF, images, …)
`scholar`	Literature search and PDF management (CrossRef, OpenAlex, Semantic Scholar)
`writer`	LaTeX manuscript compilation, BibTeX management, Overleaf export
`audio`	Text-to-speech and audio utilities (scitex-audio)
`ai`	LLM APIs (OpenAI, Anthropic, Google, Groq) and ML tools (scikit-learn)
`browser`	Web automation via Playwright
`capture`	Screenshot capture (mss, Playwright)
`dataset`	Scientific dataset access (DANDI, OpenNeuro, PhysioNet)
`cloud`	Cloud integration utilities
`app`	Unified file storage SDK (scitex-app)
`session`	Session decorator with reproducibility logging
`diagram`	Diagram generation (Mermaid, Graphviz)
`db`	Database access (SQLAlchemy, PostgreSQL)
`cv`	Computer vision (OpenCV, Pillow)
`dsp`	Digital signal processing (scipy, tensorpac)
`social`	Social media posting (socialia)
`tunnel`	SSH tunnel management (scitex-tunnel)
`all`	Everything above

Example combinations:

# Neuroscience analysis
pip install "scitex[plt,stats,dsp,dataset]"

# Paper writing
pip install "scitex[writer,scholar,plt]"

# AI agent development
pip install "scitex[ai,browser,capture]"

Development Install

Clone the repository and install in editable mode with development tools:

git clone https://github.com/ywatanabe1989/scitex-python.git
cd scitex-python
pip install -e ".[dev]"

The dev extra includes pytest, ruff, mypy, Sphinx, and build tools.

Using uv (recommended)

uv resolves and installs dependencies in parallel from a Rust resolver — roughly 10–30× faster than pip on scitex[all] (3 min vs. 30–90 min). Install it once with pip install uv (or the standalone shell installer), then prefix every install command on this page with uv:

uv pip install -e ".[dev]"

Verifying the Installation

import scitex as stx
print(stx.__version__)

To check which optional modules are available:

stx.usage.list()