Pyodide Dives

Browser Python, one experiment at a time.

A small lab for seeing where Pyodide feels natural: running code, loading packages, touching files, and passing results back to the page.

Runtime

Loading Pyodide in this browser.

Python
...
Pyodide
...
Startup
...
Packages
none
01 Python scratchpad Run snippets and inspect results. 02 CSV notebook-lite Summarize a CSV with pandas in Pyodide. 03 Python model machine specs Build a dashboard model in Python, then render it crisply on canvas.
01

Python scratchpad

Execute Python in the browser and watch stdout, errors, timing, and return values cross back into JavaScript.

Packages used No non-standard Python packages
Waiting for Pyodide. 
Output will appear here.
02

CSV notebook-lite

Drop in a CSV and let pandas, loaded inside Pyodide, produce a compact summary without sending the file to a server.

Packages used pandas
Waiting for a file. 
CSV summary will appear here.
03

Python model machine specs

Collect the machine details that browsers expose, pass them into Python for normalization, and render the dashboard with a crisp browser canvas.

Browser sandboxing hides exact CPU, RAM, disk, and GPU details, and the hints shown here may be incomplete or unreliable because browsers do not expose every value consistently.

Packages used No non-standard Python packages
Waiting for Pyodide. 
Specs render when Pyodide is ready.

Runtime probes

Bounded probes only: adjust the caps, but the app still clamps them so the browser can report useful Pyodide behavior without hunting for crash limits. If the runtime stops responding, reset the page.

Waiting for Pyodide.
Probe results will appear here.