Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

marimo

By marimo · marimo.io

  • Free plan
  • Windows
  • macOS
  • Linux
  • Web
  • Browser extension
  • API
  • Self-hosted
8.3 Club score Visit marimo.io

What marimo is

marimo is an open-source reactive Python notebook for data work, reproducible experiments, and building apps. When a cell runs, marimo runs dependent cells or marks them stale to keep code and outputs consistent. Notebooks are stored as Python files that can be tracked with Git, run as scripts, or reused as modules. Users can add sliders, dropdowns, interactive dataframe tables, and selectable plots, then deploy notebooks as interactive web apps or run them in a browser through WebAssembly. SQL cells work with dataframes, databases, CSVs, and Google Sheets, with named backends including Polars, Pandas, PyArrow, DuckDB, SQLite, and Postgres. AI notebook features support OpenAI, Anthropic, Gemini, Ollama, and OpenAI-compatible APIs. Editor extensions are available for VS Code and PyCharm; the VS Code extension also works with Cursor and other VS Code forks. The open-source notebook is free. Its hosted option, molab, is also free while usage is reasonable, with notebooks public by default.

Who it suits

marimo suits data scientists, analysts, educators, AI and analytics engineers, and web developers. It is also relevant to people who want notebooks they can version with Git or deploy as interactive apps.

Good to know, and what to watch

  • Reactive execution keeps dependent cells and outputs aligned
  • Notebooks are Python files that work with Git
  • Can deploy notebooks as interactive web apps
  • SQL supports multiple data sources and backends
  • Free and open source
  • molab notebooks are public by default
  • Running notebook code can execute arbitrary code
  • No version history is listed
  • molab saves no files larger than 1 GB after shutdown for notebooks created after August 26, 2026

marimo cloud

—

marimo open-source notebook

Free

molab

Free

Free plan

Getting started

  1. Install marimo with pip, uv, Pixi, or conda
  2. Use the open-source notebook locally or run it self-hosted
  3. Connect a VS Code or PyCharm integration if desired
  4. Create a notebook with Python or SQL cells
  5. Add sliders, dropdowns, tables, or selectable plots
  6. Deploy the notebook as a web app, script, or browser-executed project

Limits to check

The free molab plan includes 4 CPUs, 32 GB RAM, and sessions of up to 12 hours, with notebooks public by default. For notebooks created after August 26, 2026, files larger than 1 GB are not saved when molab shuts down.

The Club's review

marimo is a reactive Python notebook for reproducible analysis, experimentation, and app building. It is best suited to data and machine-learning practitioners, analysts, educators, and developers who want notebook work to behave like maintainable Python projects. The local open-source product is the strongest proposition; hosted use is convenient but has important privacy and storage trade-offs.

Overview

marimo keeps notebooks close to ordinary software development. Each notebook is a pure Python file that can be committed to Git, executed as a script, or imported as a module, while still supporting the exploratory workflow people expect from notebooks. Python and SQL are the supported languages, so teams standardized on Python can move between experiments, production code, and shareable applications without changing formats.

Deployment options extend beyond an editor: notebooks can run as interactive web apps, as Python scripts, or in browsers through WebAssembly. That makes marimo a credible choice for turning analysis into a small usable tool, not merely documenting a result.

Key features

Reactive execution recalculates dependent cells when a cell changes, or marks them stale when they cannot be run immediately. Combined with deterministic execution and package requirements stored in the notebook file, this reduces the hidden-state and out-of-order-output problems common in conventional notebooks. It improves reproducibility, although executing any notebook still deserves a code review first.

Built-in sliders, dropdowns, interactive dataframe tables, and selectable plots let users explore parameters and results without rewriting cells. SQL cells can query dataframes, CSV files, Google Sheets, and databases, with integrations including Polars, Pandas, PyArrow, DuckDB, SQLite, Postgres, and MySQL. This is a practical fit for analysts who alternate between Python transformations and SQL investigation.

AI notebook features connect to OpenAI, Anthropic, Gemini, Ollama, and OpenAI-compatible APIs. Developers can work through extensions for VS Code and PyCharm; the VS Code extension also works in Cursor and other VS Code forks. Project environments can use pip, uv, Poetry, Pixi, or Rye, while notebook sandboxes support uv and Pixi. GPU compute is supported.

Security controls are sensible but not a substitute for trust. marimo says code does not run without explicit user action, and edit-mode notebooks are loaded statically rather than executed as Python modules. Token authentication is enabled by default for remote edit-mode access, with token flags or custom middleware for run mode. Running notebook code can execute arbitrary code, so untrusted notebooks should be inspected before execution.

Pricing

marimo open-source notebook — free. This is free, permissively licensed software for local or self-managed use, and is the right plan for individuals and teams that can provide their own environment.

molab — free. The hosted notebook includes 4 CPUs, 32 GB of RAM, and sessions lasting up to 12 hours while usage remains reasonable. Notebooks are public by default, making it suitable for public teaching, demos, and sharing rather than confidential work. For notebooks created after August 26, 2026, files larger than 1 GB are not saved when a session shuts down.

marimo cloud — custom pricing. It is in closed beta and offers on-demand cloud resources for experimentation, collaboration, sharing, and deployment. Organizations needing managed collaboration should enquire rather than assume the free hosted limits apply.

Platforms

marimo is available through its API, web, Linux, macOS, and Windows, with self-hosted deployment and editor extensions. Its browser execution through WebAssembly is useful for distributing interactive work without requiring a local Python installation.

Who it's for

The maker targets AI engineers, analysts, educators, analytics engineers, data engineers, data scientists, ML engineers, and web developers. It is especially compelling for Python teams that need Git-friendly notebooks, SQL access, and a path from analysis to an interactive app. It is less suitable for groups whose primary language is not Python or who require private hosted notebooks on the no-cost tier.

Pros and cons

  • Pro: Reactive, deterministic execution helps keep outputs aligned with changed code.
  • Pro: Plain Python files work naturally with Git, scripts, and modules.
  • Pro: SQL, data-source integrations, controls, and app deployment cover a broad analysis-to-delivery workflow.
  • Pro: Free local software and self-hosting avoid a mandatory service subscription.
  • Con: molab notebooks are public by default, limiting its fit for private data.
  • Con: molab sessions stop after 12 hours and large files over 1 GB are not saved after shutdown for newer notebooks.
  • Con: Executing notebooks can run arbitrary code, so convenience does not remove security review.

Alternatives

Choose Deepnote instead if you want a freemium hosted workspace with up to three editors, five projects, and unlimited Basic machines on its free plan.

Datalore is a better fit when a hosted free tier with 120 CPU S machine hours, 10 GB of cloud storage, and two concurrent notebooks matches your workload.

Pick Hex for a collaborative analytics product whose Community plan connects data sources, supports all cell types, and includes small compute.

Nextjournal suits users prioritizing unlimited private notebooks and managed data and secrets features on its free plan.

Choose Paperspace Notebooks when included storage and free CPU and GPU machines are the deciding factors.

JupyterHub is the straightforward choice for open-source, self-hosted notebook service on your own hardware or cloud.

Pick JupyterLab for a free, open-source notebook interface installed locally with pip.

Apache Zeppelin is an alternative open-source notebook under the Apache License 2.0.

For broader discovery, browse Notebook Software and Coding playgrounds.

Verdict

Choose marimo if you want Python notebooks that remain reproducible, reviewable in Git, and deployable as interactive apps, with SQL and modern data tooling included. Look elsewhere when private hosted notebooks, longer-running sessions, or a non-Python-first workflow matters more than its unusually strong software-engineering model.

The verdict

marimo combines reactive notebooks with a Python-file format, interactive controls, SQL, and app deployment. The free hosted option makes notebooks public by default, and users should review code from untrusted sources before running it.

Questions people ask

Is marimo free and open source?

Yes. The open-source notebook plan is free and permissively licensed.

Which languages and formats are supported?

The supported languages are Python and SQL. Notebook formats include Python, Jupyter Notebook, HTML, PDF, and Markdown.

Can marimo run without a hosted service?

Yes. It can run on Linux, macOS, and Windows, and supports self-hosted deployment.

What is molab?

molab is a free cloud-hosted notebook option with 4 CPUs, 32 GB RAM, sessions up to 12 hours, and notebooks public by default.

Which AI services can connect?

AI notebook features support OpenAI, Anthropic, Gemini, Ollama, and OpenAI-compatible APIs.

Is running notebook code safe?

Running code can execute arbitrary code. marimo advises using trusted notebooks or reviewing their code first.

What the maker publishes

AI integrations
The AI notebook features support OpenAI, Anthropic, Gemini, Ollama, and OpenAI-compatible APIs. source
Authentication
Token authentication is enabled by default for remote access in edit mode; run mode supports token flags or custom middleware. source
Company
The site identifies the company as Marimo Inc. and names Akshay Agrawal as co-founder and CEO; the opened pages do not state headquarters or founding year. source
Deployment
Notebooks can be deployed as interactive web apps, run as Python scripts, or executed in browsers via WebAssembly. source
Deployment
Notebooks can be served as web apps, executed as scripts, or exported as WebAssembly-powered HTML. source
Developer integrations
marimo offers extensions for VS Code and PyCharm, and its VS Code extension also works in Cursor and other VS Code forks. source
Editor integrations
The documentation describes a VS Code extension that works with VS Code, Cursor, and other VS Code forks, plus a PyCharm plugin. source
Founded
2024 source
Free hosted option
molab is described as a free cloud-hosted notebook platform, with notebooks public by default. source
Free plan
Yes source
Git-friendly format
Notebooks are stored as pure Python files that can be versioned with Git, run as scripts, and reused as modules. source
Install options
The installation guide lists pip, uv, Pixi, and conda as package manager options. source
Intended users
The maker lists AI engineers, analysts, educators, analytics engineers, data engineers, data scientists, ML engineers, and web developers as user groups. source
Interactive features
Built-in UI elements include sliders and dropdowns, and the editor includes interactive dataframe tables and selectable plots. source
Interactive tools
marimo includes interactive elements such as sliders, dropdowns, dataframe tables, and selectable plots. source
molab storage limit
For notebooks created after August 26, 2026, files larger than 1 GB are not saved when the notebook shuts down. source
Package management
For project package management, marimo supports pip, uv, Poetry, Pixi, and Rye; notebook sandboxes support uv and Pixi. source
Product
marimo is an open-source reactive Python notebook that automatically runs dependent cells and stores notebooks as Git-friendly Python files. source
Reactive execution
When a cell runs, marimo runs dependent cells or marks them stale to keep code and outputs consistent. source
Reproducibility
marimo uses deterministic execution without hidden state and can store package requirements in notebook files. source
Security
marimo says no user code runs without explicit user action, and notebooks opened in edit mode are statically loaded rather than executed as Python modules. source
Security and subprocessors
The maker says subprocessors must provide security and confidentiality protections consistent with its customer obligations and lists CoreWeave, Cloudflare, and Vercel among them. source
Security limitation
Running notebook code can execute arbitrary code, so marimo advises running notebooks from trusted sources or reviewing their code first. source
SQL and data
SQL cells work with dataframes, databases, CSVs, and Google Sheets; the home page names Polars, Pandas, PyArrow, DuckDB, SQLite, and Postgres among supported backends. source

Checked against the maker's pages on 7 Oct 2026.