marimo is an open-source notebook for Python and SQL, designed from the ground up for working with data — run reproducible experiments, version with Git, share as interactive web apps, and execute as Python scripts. All in a modern, AI-powered editor.

👋 Hello, Product Hunt! My name is Akshay, and along with my co-founder Myles (Palantir, CloudKitchens) I'm excited to share marimo, an open-source notebook for Python and SQL. Unlike Jupyter notebooks, marimo notebooks are reproducible, git-friendly (stored as pure Python), shareable as interactive web apps, and executable as scripts. When I was a PhD student in machine learning at Stanford and an engineer at Google Brain, I relied on Jupyter notebooks to run experiments and prototype algorithms because they let me see my data while I worked on it. But they were also extremely frustrating because they aren't reliably reproducible (over a third of the 10 million notebooks on GitHub failed to reproduce!), they're stored as JSON (rather than Python), and the final document is static with no interactivity. marimo solves these problems by blending the best parts of interactive computing with the rigor of traditional software development. Think of marimo as a modern replacement for Jupyter, Streamlit, and Gradio, with these key features: 🔬 Reproducible: marimo keeps code and outputs in sync, eliminating hidden state. Built-in package management means that notebook files can document their own dependencies. 🐍 Git-friendly: Notebooks are stored as .py files, letting you version them with Git. 🖐️ Interactive: Bring data to life with interactive elements. Search, filter, and sort dataframes without code; select points in a plot and get them back as dataframes; bind sliders, dropdowns, and more in just a line of code. 🛜 Deploy as web apps: Deploy notebooks as an interactive web apps or slides with a single command. 🏃 Execute as scripts: Execute any notebook as a Python script from the command line. 🛢️ Designed for data: Query dataframes and databases with SQL, and see all your data sources in a side panel Transform dataframes using our no-code dataframe editor. 🤖 AI built-in: Code in a modern editor that supports GitHub Copilot and AI assistants; generate code that refers to the schemas of your dataframes. Get started To get started, run: pip install marimo && marimo tutorial intro To get inspired, view our community gallery: https://marimo.io/gallery To help, leave a star: https://github.com/marimo-team Our inspiration We believe that the tools we use shape the way we think — better tools, for better minds. Our inspiration comes from many places and projects, especially Pluto.jl, ObservableHQ, and of course Jupyter, which established interactive computing as a pillar of computational science through open tools and standards. marimo is part of a greater movement toward reactive dataflow programming. From IPyflow, streamlit, TensorFlow, PyTorch, JAX, and React, the ideas of functional, declarative, and reactive programming are transforming a broad range of tools for the better. Finally, big thanks to our community and all the hunters here for your support, as well as a massive thanks to @chrismessina, our hunter and our shepherd, who believed in marimo's mission and helped us launch here. 🙏🙇

I’ve been using Marimo for over a year and it is the best damn notebook environment I’ve ever used. When I was learning ML, half my time was spent debugging variables that were out of sync with my cells, figuring out the commands to plot things correctly, and other annoying tedium. No more. I make stuff in Marimo and the slider elements and dynamically updating cells are soooooo cool. I didn’t realize notebooks could be fun! This is so obviously the future of data science. Looking forward to see what else Marimo team builds going forward.

We are using marimo in production at Owkin (AI biotech), to build data visualization apps. To me it is really the fastest tool to build simple apps, 1 order of magnitude faster than concurrents like streamlit. We do simple prototypes, and then iteratively grow them into production grade apps. Secondly, the intuitive notebook-based approach + the interop with rich python ecosystem, enables our data scientists, computational biologists, and more, to contribute to app building. Last, since it is all defined as code (git friendly), we can easily manage contributions from all stakeholders: data preparation code, plots, components, apps. It's awesome !

This is the best Python notebook! I'm a power user of notebooks and none of them come close to the reactivity and interactivity marimo offers. I also really like that the file format is pure Python. I can't wait to see what's next.

marimo

The next-generation Python notebook

The next-generation Python notebook

Have a question about marimo? Ask it here and get a real answer.

marimo Launches

Do you use marimo?

Maker Comment

Have a question about marimo? Ask it here and get a real answer.

marimo Launches

Do you use marimo?

Maker Comment