This is your monthly reminder that at marimo, we’re on a mission to make the
world’s best development environment for AI and data, and it’s all free
and open source. Just pip install marimo or uv add marimo to get started!
You’re reading the 19th marimo newsletter. In this newsletter, we’ll give highlight some new features, share some exciting case studies of how companies use marimo, and last but not least share some big news (make sure to read until the end!).
New features
In the last few weeks, we’ve added many new features and many enhancements. Here are just a few highlights:
- OpenAI Codex support. Use OpenAI’s Codex agent in the AI chat sidebar by selecting “Codex” from the agent dropdown (bring your own keys). The agent can help with code editing, debugging, and notebook modifications.
- Altair and Vega-Lite Enhancements marimo now uses Vega-Lite 6. Charts support legend-based selection in composite charts (layered, vconcat, hconcat), binned data selection, and images in tooltips. Table charts are now faster and handle string columns with up to 20k rows.
- Experimental tools in the chat sidebar. When enabled, the chat sidebar can run with tools to edit cells, run cells, and read their outputs, using model provider (or local model) of your choice.
See our changelog for the full list: https://github.com/marimo-team/marimo/releases
Case studies
Why Norway’s largest bank chose marimo over Databricks notebooks
DNB’s Cyber Defense team is replacing Databricks with a composable security data platform built on marimo, Ibis, and Iceberg.
As Norway’s largest financial institution, DNB processes millions of security events daily. Their team needed a flexible interface that could unify investigation, visualization, and documentation, without vendor lock-in or distributed compute overhead. marimo was the obvious chose to be the unified interface and AI layer for their data platform, powering faster and more systematic investigations across multiple backends.
Read the full story by DNB Staff Engineer Kyrre Wahl Kongsgård here: https://marimo.io/blog/case-study-dnb
Why everything about Sumble lives in marimo
Sumble (founded by Anthony Goldbloom, founder of Kaggle) replaced Jupyter with marimo across their entire company.
As a fast-growing team building AI-powered account intelligence tools, they needed a better way to turn notebooks and data insights into dynamic internal tools and dashboards. Today, marimo is the single source of truth for their entire organization, seamlessly powering 25+ internal applications and dashboards, unifying exploration, collaboration, and deployment in one environment.
Learn why Anthony says marimo has been “truly transformational” for his company in the case study: https://marimo.io/blog/case-study-sumble
How marimo enables radiologists, MLEs, and backend at Bunkerhill Health
Bunkerhill Health, a Sequoia and Felicis-backed Series A startup, replaced Jupyter notebooks with marimo across their ML, backend, and clinician workflows.
As a fast-growing AI startup, Bunkerhill’s ML and backend engineers work with millions of image-text pairs to build AI agents that support clinicians. Previously, that work spanned fragmented notebooks, dashboards, and scripts, significantly slowing iteration and deployment. With marimo, they gained a unified interface that combines development, visualization, and deployment in one environment.
Read the full case study: https://marimo.io/blog/case-study-bunkerhill
Marimo joins CoreWeave
Finally, we have some big news to share: Marimo is joining CoreWeave! We’re doubling down on open-source and scaling molab with serious compute.
Read the full announcement at our blog: https://marimo.io/blog/joining-coreweave.
Our mission remains the same: to build the world’s best open-source programming environment for working with data. To our community: we wouldn’t be here without your enthusiasm, feedback, and contributions. We’re so grateful — and so excited to keep building with you.
marimo is, and always will be, free, open-source, and permissively licensed.
📬 In case you missed it …
- 🔒 We have a new security policy at our docs.
- 🛞 Learn how to use marimo in Airflow or Prefect pipelines
🍃 Community
We’re at 1M+ monthly downloads, 16k+ stars on GitHub, have nearly 200 contributors pushing code to marimo, 600k+ YouTube views, and nearly 3k+ marimonauts hanging out on Discord — join the conversation!
Roundup.
🥇 Agentic Visual Reporting. Péter Gyarmati’s won 1st place at IEEE VIS conference’s VISxGenAI workshop, with his submission on Agentic Visual Reporting, built using marimo and quak (by marimo team member Trevor Manz!).
🌌 Quasar data and marimo. Check out RealPython’s guide on investigating Quasar data with polars and marimo!
🤖 Agents and marimo check. Eric J Ma’s blog post on how to teach your coding agents with AGENTS.md, featuring a shout out to linting command marimo check.
Don’t forget to submit your projects to our awesome-marimo repo!
Sincerely,
marimo team 🍃
