Marimo Inc.
Building the world's best open-source programming environment for working with data. Marimo is an open-source reactive notebook for Python that is reproducible, Git-friendly, deployable as a web app, and executable as a script.
At a Glance
- AI/ML engineers and researchers
- Data scientists and analysts
- Computational scientists
- Academic researchers and universities
- +5 more
AI Tools by Marimo Inc.
(3)
marimo.app
Browser Python Notebook Playground
molab
Cloud Hosted Marimo Notebooks
marimo
Reactive Python Notebook IDE
Discussions
No discussions yet
Be the first to start a discussion about Marimo Inc.
Latest News
Released version 0.19.6 - latest release with 341 total releases
Newsletter 21 - Hiring, v0.19.0, 2025 recap, and 2026 preview
Announcing VS Code and Cursor extension for marimo notebooks
Marimo is Joining CoreWeave - Acquisition announcement
Products & Services
Open-source reactive Python notebook stored as pure Python files (not JSON). Provides reproducible execution, interactive UI elements, can be executed as scripts, and deployed as web apps. Features reactivity like a spreadsheet, eliminating hidden state common in Jupyter notebooks.
Cloud-hosted marimo notebook workspace that lets users rapidly experiment on data using Python and SQL. Free to use with ability to upload data files, generate code with AI, share links, and download notebooks. After CoreWeave acquisition, offering expanded compute resources including GPU-accelerated instances.
Native extension for VS Code and Cursor that brings marimo notebook experience directly into these popular code editors.
Market Position
Positions as a next-generation alternative to Jupyter notebooks, addressing long-standing problems like hidden state, non-reproducibility, and poor version control. Inspired by Pluto.jl and Observable, part of the reactive dataflow programming movement. Key differentiators include: reactive execution (automatically runs dependent cells), pure Python storage (not JSON), ability to execute as scripts and deploy as apps, no hidden state for reproducibility, and built-in SQL support. Competitive advantages over Jupyter include Git-friendliness, deterministic execution, and seamless transition from prototype to production.
Leadership
Founders
Akshay Agrawal
Former engineer at Google Brain where he worked on TensorFlow. PhD from Stanford University with research focus on vector embeddings and machine learning. Published at NeurIPS. Helped grow an open-source library for optimization (CVXPY) to a million monthly downloads.
Myles Scolnick
Former engineer at Palantir and CloudKitchens. Led teams across product and web platforms. 11 years of experience building data-first developer tools and web platforms.
Executive Team
Akshay Agrawal
Co-Founder, CEO
Former Google Brain engineer who worked on TensorFlow. PhD from Stanford University focused on vector embeddings and machine learning.
Myles Scolnick
Co-Founder, CTO
Former Palantir and CloudKitchens engineer who led teams across product and web platforms.
Board of Directors
Founding Story
Started nearly four years ago (around 2022) when Akshay Agrawal was motivated to create better tools for Python developers working with data. His experience researching and developing vector embedding methods in Jupyter notebooks, which were extremely frustrating due to hidden state, non-reproducibility, and poor version control, led him to create something better. Myles Scolnick joined as co-founder, and together they created marimo to blend the best parts of interactive computing with the rigor of traditional programming. The project was incubated for over a year with Stanford's SLAC Laboratory before being publicly launched on HackerNews in January 2024.
Business Model
Revenue Model
Open-source software with cloud-hosted notebook service (molab). Initially free to use with plans to offer paid tiers for enhanced compute resources and GPU access. After CoreWeave acquisition, revenue model includes integration with CoreWeave's cloud services.
Pricing Tiers
Core marimo notebook is free, open-source, and permissively licensed. Can be self-hosted and used without restrictions.
Cloud-hosted notebooks with reasonable usage limits. After CoreWeave acquisition, offering more generous free tier with larger instances, GPUs, and longer-lived sessions.
Planned paid tiers for GPU-accelerated compute services and enhanced resources through CoreWeave Cloud.
Target Markets
- AI/ML engineers and researchers
- Data scientists and analysts
- Computational scientists
- Academic researchers and universities
- Software developers working with data
- Educational institutions
- Reproducible scientific experiments and research
- Data science workflows involving SQL, dataframes, and visualization
- Building and deploying interactive data apps and internal tools
- Educational purposes for teaching computational science
- Collaborative development requiring Git-friendly notebook formats
- Rapid prototyping and sharing of Python experiments
- Stanford University
- Johns Hopkins University
- UC Berkeley
- Princeton University