marimo
A VS Code extension for marimo, a
reactive Python notebook that's reproducible, git-friendly, and deployable as
scripts or apps.
Prerequisites
This extension requires uv to be installed and
available on your PATH. The extension uses uvx to run the bundled
marimo-lsp language server.
Install uv:
# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
# Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
Features
- 📓 Create, edit, and share marimo notebooks in VS Code
- ✨ Rich output rendering with interactive UI elements
- 📦 Integrated package and environment management
- 🔍 Variables explorer and datasources panel
Getting Started
- Ensure uv is installed
- Install this extension from the VS Code Marketplace
- Open a marimo notebook (
.py file), or create a new one:
- Press
Cmd+Shift+P (macOS) or Ctrl+Shift+P (Windows/Linux)
- Run "Create: Marimo notebook"
- VS Code will prompt you to open the file as a marimo notebook
If you have an existing Python file that's a marimo notebook, you'll see an
icon in the editor title bar to open it as a notebook (see image above).
Commands
| Command |
Description |
marimo: New marimo notebook |
Create a new marimo notebook |
marimo: Open as marimo notebook |
Open a Python file as a marimo notebook |
marimo: Toggle on-cell-change mode |
Switch between auto-run and lazy execution modes |
marimo: Run stale cells |
Execute all cells that need to be re-run |
marimo: Restart notebook kernel |
Restart the notebook's Python kernel |
marimo: Create setup cell |
Create or navigate to an existing setup cell |
marimo: Publish notebook as Gist |
Share your notebook as a GitHub Gist |
marimo: Publish notebook as... |
Export your notebook in various formats |
marimo: Export static HTML |
Export notebook with current outputs as HTML (without re-executing cells) |
marimo: Set Python interpreter to match notebook kernel |
Set the active Python interpreter to match the notebook's kernel |
marimo: Restart marimo language server (marimo-lsp) |
Restart the LSP server if it becomes unresponsive |
marimo: Report an issue or suggest a feature |
Open GitHub to report bugs or request features |
marimo: Show marimo diagnostics |
Display diagnostic information for troubleshooting |
Configuration
| Setting |
Type |
Default |
Description |
marimo.lsp.path |
array |
[] |
Path to a custom marimo-lsp executable, e.g., ["/path/to/marimo-lsp"]. Leave empty to use the bundled version via uvx. |
marimo.uv.path |
string |
|
Path to the uv binary, e.g., /Users/me/.local/bin/uv. Leave empty to use uv from the system PATH. |
marimo.disableUvIntegration |
boolean |
false |
Disable uv integration features such as automatic package installation prompts. |
marimo.disableManagedLanguageFeatures |
boolean |
false |
Disable marimo's managed Python language features (completions, diagnostics, formatting). When enabled, notebook cells use the standard python language ID and rely on external language servers. |
marimo.telemetry |
boolean |
true |
Anonymous usage data. This helps us prioritize features for the marimo VSCode extension. |
Language Features
By default, marimo provides managed Python language features (completions, diagnostics, formatting) for notebook cells using a dedicated language server. This prevents conflicts with external Python language servers and ensures a consistent editing experience.
Why managed mode? Notebook cells use a custom language ID (mo-python) to isolate them from external language servers like Pylance or Jedi. This prevents duplicate completions, conflicting diagnostics, and other interference while editing marimo notebooks.
When to disable managed mode: If you prefer to use your own Python language server configuration (e.g., Pylance, pyright, or another LSP), enable marimo.notebook.disableManagedLanguageFeatures. This switches cells to use the standard python language ID, allowing external language servers to provide completions and diagnostics. Note that this may result in some language features not working as expected in the notebook context.
Support
License
Apache 2.0
