Live editing

Notebook preview

In Jupyter (or any frontend that honours _repr_html_), returning a deck, a slide, or a single cell as the last expression in a cell renders an inline, sandboxed preview — no need to write a file or open a browser:

deck = Deck(title="Q3 Report", plugins=[Plugins.Plotly()])
deck.add_title("Q3 Report")
slide = deck.add_slide("KPIs", nrows=1, ncols=2)
cell  = slide.add_metric(value=98.7, label="Efficiency")

deck     # full deck preview (with sidebar / toolbar)
slide    # just this slide, chrome-free
cell     # just this cell

Each preview is embedded in an isolated <iframe> so the deck’s styling and scripts never leak into the notebook. Slide and cell previews reuse the deck’s theme and plugins.

Laying out cell outputs

Because a single slide previews on its own, you can use Montin purely as a layout engine for notebook outputs — build one slide, drop a few cells onto its grid, and return it as the cell’s last expression. No file is written:

import plotly.express as px
from montin import Deck, Plugins

deck  = Deck(title="scratch", plugins=[Plugins.Plotly()])
slide = deck.add_slide("Run 42", nrows=1, ncols=3)

slide.add_metric(value=0.94, label="Accuracy", delta=+0.02)
slide.add_plotly(px.line(df, x="epoch", y="loss"))
slide.add_table(summary_df)

slide   # ← renders a laid-out mini-dashboard as the cell output

This gives you a tidy grid of metrics, charts, and tables in a single cell — handy for experiment summaries — without leaving the notebook.

Autosave

Pass autosave with a filename to have Montin write the HTML file automatically after each change, so you can live-preview in a browser while building the deck.

deck = Deck(
    title="Live Report",
    autosave="report",        # writes report.html after each trigger
    autosave_level="cell",    # trigger: "slide" (default) or "cell"
)

autosave_level controls how frequently the file is written:

Level	Trigger	Recommended when
"slide"	After each add_slide()	Large reports with many cells
"cell"	After each add_*() cell call	Small decks or iterative Jupyter workflows

Note

For large reports prefer autosave_level="slide" to avoid redundant writes on every cell addition, or keep autosave disabled, manually saving with .write() when needed.

VSCode workflow

A convenient way to use autosave is alongside the Live Preview extension in VSCode. Open the generated .html file with Live Preview (right-click the file → Show Preview) and it will automatically refresh whenever Montin rewrites it. This gives you an instant side-by-side view — code on the left, the rendered report on the right — without leaving the editor, as shown below: