Clear Jupyter Notebook Output Online (.ipynb Cleaner)

The Notebook Output Cleaner removes the saved outputs and execution counts from a Jupyter .ipynb file, giving you a clean version that's ideal for version control. Outputs — printed text, tables, and especially embedded images and charts — are stored inside the notebook JSON as large base64 blobs, so committing a notebook with its outputs produces enormous, noisy git diffs and bloats your repository. Clearing them first keeps history readable and files small.

It does the same job as the command-line nbstripout tool, but in your browser with nothing to install. Paste or drop your .ipynb, and it sets every code cell's outputs to empty and its execution_count to null, then hands back a cleaned notebook you can copy or download. Optionally it also removes run-state metadata (cell timing, collapsed/scrolled flags) and interactive widget state, which are the other common sources of noisy diffs.

Crucially, the cleaner never touches your code: markdown cells and the source of every code cell are preserved exactly. It only removes the generated results, which you regenerate simply by re-running the notebook. The output is a valid .ipynb that opens normally in Jupyter, JupyterLab, VS Code, or Colab.

Everything runs in your browser — the notebook is parsed and cleaned client-side. If you want to read a notebook rather than clean it, use the companion Notebook Viewer; to extract the code as a .py script, use the Jupyter-to-Script converter.