Python Notebooks Without the Fuss.
UV-native.
Dependencies just work. No kernel configuration, no environment headaches.
SQL is first-class.
Query DataFrames with SQL, or mix Python and SQL freely.
pynb works with Codex, with more agents coming soon.
Use your ChatGPT subscription.
Git-friendly.
Cells marked with # %% comments. Diffable, no merge conflicts.
Lazy by default.
Work with datasets larger than memory.
macOS only for now.
Nope. pynb uses uv to handle your Python environment. If you already have uv installed, pynb will use it.
It's a different implementation of Jupyter notebooks, focused on simplicity and local-first workflows.
It's a percent format used by many Python tools. Cells are marked with # %% comments. The benefit is that it diffs well in git.
Yes.
It can manage and run notebooks on your behalf.
You need one for the AI features.
Yes, Claude support is coming soon.
No. Your data stays on your machine. We only collect anonymized usage data (nothing sensitive).
Locally on your machine. We don't store any of your data.