figstudio 0.4.0

A local figure-workflow tool for reproducible, publication-ready Matplotlib panels.

FigStudio

FigStudio is a local-first figure workflow for scientific Python users who want to turn data already present in a script or notebook into polished, reproducible Matplotlib panels.

It opens a browser editor from your Python process, lets you map live variables to plot layers or statistics recipes, create faceted DataFrame panels, add reference lines and annotations, previews through Matplotlib, exports publication files, and saves plain Matplotlib OO code back to a controlled script block or notebook cell.

Install

pip install figstudio

The wheel includes the React editor. End users do not need Node, npm, Vite, or the frontend source tree after installation.

Try It

figstudio demo

For a script, add FigStudio after data preparation and give generated plotting code one controlled block:

import figstudio

# Prepare data first.
session = figstudio.open(locals(), script_path=__file__, block_id="main")

# figstudio:start main
# figstudio:end main

In the editor, choose a live variable, add a plot layer or statistics recipe, polish the figure, export PNG/SVG/PDF, then click Save code.

Documentation

Language	Start here
English	docs/en/index.md
Chinese / 中文	docs/zh/index.md

Common entry points:

Reader	English	中文
Figure users	Get Started	快速开始
Scientific workflows	Workflows	科研制图工作流
Example gallery	Gallery	Gallery
API consumers	API Reference	API 参考
Contributors	Developer Guide	开发者指南
Product scope	PRD	产品需求
Roadmap overview	Roadmap	路线图
Roadmap strategy	Strategy	路线图策略
Roadmap initiatives	Initiatives	路线图 Initiatives
Deferred roadmap items	Deferred Work	暂缓事项
Release history	Release Notes / CHANGELOG	发布说明 / CHANGELOG

Development

uv run --extra dev pytest
cd frontend
npm install
npm run build
npm run dev

Build a publishable package:

uv build

The build hook bundles the frontend into the Python wheel. Runtime installs from a built wheel still use the packaged editor and do not require frontend tooling.