ipynb to Script

What this tool does

Paste a Jupyter notebook — the full contents of an .ipynb file, which is just JSON — and get back a clean, runnable Python script. It runs entirely in your browser with WebAssembly: nothing is uploaded, it works offline, and there's no sign-up.

The converter:

• extracts every code cell in order, verbatim,
• drops all outputs and execution counts (In[3]:, stored stdout, plots, tracebacks) so you get source, not a run log,
• turns markdown cells into # comments (or keeps them verbatim in Markdown mode) so the narrative survives,
• skips empty cells so the script stays tidy,
• understands both nbformat v4 (top-level cells) and older v3 (worksheets), and
• reads the notebook's own language metadata for Markdown code fences (defaults to python).

Options

Worked example

Given this notebook (two cells — one markdown, one code with a stored output):

{
  "cells": [
    {"cell_type": "markdown", "source": ["# Demo notebook\n", "\n", "Load and greet."]},
    {"cell_type": "code", "execution_count": 1,
     "outputs": [{"output_type": "stream", "name": "stdout", "text": ["hi\n"]}],
     "source": ["import sys\n", "print('hi')"]}
  ],
  "metadata": {"language_info": {"name": "python"}},
  "nbformat": 4, "nbformat_minor": 5
}

the default script output is:

# # Demo notebook
#
# Load and greet.

import sys
print('hi')

The execution_count and the stored hi output are gone; the markdown cell became # comments. Turn on Add # %% cell markers and each cell is prefixed with # %% (# %% [markdown] for prose) so VS Code shows cell boundaries. Turn on Include stored cell outputs and a # Output: / # hi comment is appended after the code.

Limits & edge cases

• Paste the JSON, not a file. A .ipynb is a JSON text file — open it and paste its contents. There's no binary upload; if the text isn't valid JSON you get a clear "input is not valid JSON" error.
• It never executes the notebook. There's no Python kernel here — it only reads outputs already stored in the file, and by default drops them. It can't produce fresh outputs.
• Empty cells are skipped, and a notebook with no convertible cells returns a clear error rather than an empty file.
• Blank line between cells: cells are joined with one blank line. Cell-internal blank lines are preserved as written.
• The output has no trailing newline — copy or download it as-is.
• # In[N]: markers and the shebang are intentionally omitted (nbconvert adds them) to keep the script clean and diff-friendly.

FAQ