NumPy and pandas without a notebook server
Write import numpy as np or import pandas as pd and press Run. The packages download once from the Pyodide CDN and then execute on your own CPU, so your data never leaves the page.
Because everything runs in the browser tab, very large datasets are limited by your device's memory. For quick calculations, checking a formula or teaching, it is faster than starting a notebook.
Frequently asked questions
Which packages are available?
Everything bundled with Pyodide, including numpy, pandas, scipy, scikit-learn, statsmodels and matplotlib. They load automatically when imported.
Can I plot charts?
Text output is supported today. Rendering matplotlib figures is on the roadmap.
Related tools
- Pine compilerCheck, run and chart v4 · v5 · v6 scripts in your browser — nothing to install, nothing uploaded.
- PythonPython 3 via WebAssembly — runs on your device.
- Python + stdininput() reads from the stdin box — one value per line.
- JavaScriptSandboxed Web Worker with top-level await.
- TypeScriptTranspiled by the real TypeScript compiler, then executed.
- LuaLua 5.3 on the Fengari VM.
- SQL (SQLite)A real SQLite engine in WebAssembly — fresh database every run.
- HTML / CSS / JSLive preview in a sandboxed frame.