marimo
A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.
marimo is an open-source dataframes and notebooks project on GitHub under the Apache-2.0 license (permissive), with 23k stars. It is built with CPython, TypeScript, NumPy and pandas. Its last push was today.
- Stars
- 23k
- License
- Apache-2.0
- Activity
- Active
- Latest release
- 0.25.2
Replaces Jupyter Notebook: why marimo is the better choice, and when it is not.
Judgement
- What it is
- Developer tool
- What it does
- A reactive Python notebook that reruns dependent cells automatically, stores notebooks as plain .py files, and can run them as scripts or serve them as web apps.
- Adopt it when
- Notebooks need to live in git, rerun reproducibly and double as small internal data apps without moving to Streamlit.
- Watch out
- It is still pre-1.0, and the reactive model rules out patterns Jupyter users rely on, such as redefining a variable in a later cell.
- Ready for
- Usable
- Runs as
- Command line and self-hosted server
- Built with
- Python and TypeScript
- How established
- Established: started 3 years ago, latest release 0.25.2.
- What the license allows
- You can build it into a product you sell, as long as you keep its copyright notice.
Judged on 2026-11-06 from the project's README and GitHub data. The license and age lines are worked out directly.
Built from
Follow it down: marimo → CPython → GCC → LLVM project → runs on a microchip
Open-source alternatives to marimo
-
1
pandas pandas-dev
Flexible and powerful data analysis / manipulation library for Python, providing labeled data structures similar to R data.frame objects, statistical functions, and much more
Building block Production-ready You analyse spreadsheets, CSVs or database extracts in Python.
★ 50k BSD-3-Clause Python Active, pushed today
-
2
DuckDB duckdb
DuckDB is a high-performance analytical database system
Building block Production-ready You analyse data that fits on one machine and want fast SQL without running a server.
★ 43k MIT C++ Active, pushed 6 days ago -
3
Polars pola-rs
Extremely fast Query Engine for DataFrames, written in Rust
Building block Production-ready Your pandas workloads are slow or run out of memory on a single machine and you can rewrite them in Polars' expression API.
★ 40k MIT Rust Active, pushed today
-
4
PyGWalker Kanaries
PyGWalker: Turn your dataframe into an interactive UI for visual analysis
Developer tool Usable Data analysts want to explore a DataFrame visually without writing plotting code.
★ 16k Apache-2.0 Python Active, pushed 2 months ago -
5
JupyterLab jupyterlab
JupyterLab computational environment.
Developer tool Production-ready Data teams work in notebooks all day and want several notebooks, consoles and files open side by side, with extensions installed from PyPI.
★ 15k BSD-3-Clause TypeScript Active, pushed today -
6
Jupyter Notebook jupyter
Jupyter Interactive Notebook
Developer tool Production-ready Analysts or researchers want the familiar single-document notebook interface for exploring data and sharing worked examples.
★ 13k BSD-3-Clause Jupyter Notebook Active, pushed 6 days ago
-
7
cuDF NVIDIA
cuDF - GPU DataFrame Library
Building block Production-ready You process large tabular data and have NVIDIA GPUs available, especially to speed up existing pandas or Polars code.
★ 9.8k Apache-2.0 C++ Active, pushed today -
8
DataFusion apache
Apache DataFusion SQL Query Engine
Building block Production-ready You are building your own database, query service or data tool and want an embeddable query engine.
★ 9.6k Apache-2.0 Rust Active, pushed today
Stars, license and activity come from GitHub and refresh on a rolling hourly schedule. The project's own description is quoted as written.