# Getting Started (/scripting-v2/notebook/getting-started)



## Open the project [#open-the-project]

Run Market Lab from the folder where notebooks and research files should live:

```bash
cd my-research
mlab notebook
```

JupyterLab opens with the current folder as its root.

## Python environment [#python-environment]

Market Lab selects the notebook's Python interpreter in this order:

1. The path supplied with `--python`.
2. `./.venv` in the current folder.
3. `python3` or `python` from `PATH`.

Select an environment explicitly when needed:

```bash
mlab notebook --python .venv/bin/python
```

The selected Python environment must contain `ipykernel`, and `jupyter-lab` must be available in that environment or on `PATH`. If either is missing, Market Lab detects the project's package manager and prints one installation command using either `uv` or `pip`.

NumPy, pandas, Matplotlib, statsmodels, and other research packages come from the selected environment. Market Lab does not install or bundle them.

## Start without opening a browser [#start-without-opening-a-browser]

```bash
mlab notebook --no-browser
```

Jupyter prints its local URL so it can be opened manually.

## Confirm the session [#confirm-the-session]

Create or open a notebook with the Market Lab kernel. The `mlab` object is already available:

```python
mlab
```

Expected output:

```text
MarketLabNotebook(connected=True, mode='research')
```

Do not run `import mlab`. Market Lab injects the object when the kernel starts; it is not an installed Python module.

Next, follow the [Research Workflow](/scripting-v2/notebook/research).