Questions about converting notebooks
- Is my notebook uploaded to a server?
- No. The file is read with the browser's file API and rendered by JavaScript on your machine. There is no upload endpoint, so the notebook cannot leave your computer — which also means you can use this on work that is not allowed to go to a third-party service. One honest caveat: if a Markdown cell links an image hosted on the web, your browser fetches that image from wherever it lives, exactly as Jupyter itself does. Images stored inside the notebook never cause a request.
- Do I need Python, LaTeX or nbconvert installed?
- No. Nothing is installed and nothing is executed. The notebook is treated as a document: its cells and their saved outputs are rendered as they were when you last ran them.
- Why does Download PDF open a print dialog?
- Because your browser's print engine is the best PDF writer available without a server. It produces real, selectable, searchable text and sane page breaks. The alternative — screenshotting the page into an image-based PDF — gives you a file where you cannot select a single line of code. Choose 'Save as PDF' as the destination and the file is written straight to disk.
- Can I export the notebook without code cells?
- Yes. Turn off the Code toggle and the PDF contains only Markdown and results. This is the usual choice when the reader cares about findings rather than implementation.
- Are LaTeX equations supported?
- Yes. Inline math with $…$ or \(…\) and display math with $$…$$ or \[…\] are rendered with KaTeX. Dollar signs inside code blocks are left alone, so shell snippets are not mistaken for equations.
- What about plots and DataFrames?
- Anything the notebook saved as output is rendered: matplotlib PNGs and SVGs, pandas DataFrame tables, stdout, stderr and error tracebacks. Outputs are read from the file, so a cell that was never run has nothing to show.
- What happens to interactive widgets and Plotly charts?
- Anything that needs JavaScript to draw itself cannot exist in a PDF — this is true of every notebook-to-PDF tool, including nbconvert. If a static image was saved alongside the widget, that image is used. The fix is to save such figures as static images in the notebook before exporting.
- How large a notebook can it handle?
- The limit is your browser's memory rather than a server quota. Notebooks with hundreds of embedded images are the slow case, since every image is base64 data inside the file itself.
- Is it really free?
- Yes — no account, no file limit, no watermark. There is no server doing the conversion, so there is no per-file cost to pass on.
Something not answered here? The comparison of the five export routes and the nbconvert error guide cover the command-line side.