Features
Reactive execution: cells update on dependency change
Git-friendly plain Python (.py) storage
Built-in SQL cells: DuckDB, Postgres, MySQL, SQLite
Interactive UI widgets: sliders, dropdowns, tables, plots
AI autocomplete, chat, and error fixing
marimo pair: connect AI agents to running sessions
Deploy notebooks as WASM HTML or web apps
Run notebooks as Python scripts via CLI
Reproducible deterministic execution, no hidden state
Package requirements embedded in notebook file
VS Code and Cursor extension for IDE integration
MCP support for AI agent connectivity
JupyterHub extension for enterprise deployment
Customizable UI with themes and vim keybindings
Debugging panels and profiler
Natural language AI writes Python, SQL, formulas, charts
Live data connectors: Postgres, MySQL, Snowflake, BigQuery, QuickBooks, Plaid
MCP support for AI agents (Claude, Cursor, ChatGPT)
API for reading/writing cells and triggering AI runs
Import CSV, Excel, PDF, Google Sheets
Scheduled tasks for automated data refresh
Conditional formatting and merged cells
STOCKHISTORY function for financial data
Images and embeds in spreadsheets
AI-generated code with explanations that are editable
Collaborative sharing with repeatable insights
Folders for organizing files
Python, SQL, JavaScript, and formula cells
Agent Sync to link apps and APIs into spreadsheets
SOC 2 and HIPAA certified
