Python Runner

Most of the work here is a Python package with no interface, and the fastest way to show what one does is to let you run it. This is CPython compiled to WebAssembly with a package installer attached, so a wheel from PyPI runs in the same tab you are reading this in.

Downloads the Python runtime and any packages from public CDNs on first use. Your code runs in this tab and is never sent anywhere.

Environment

Not started

The standard library, and any pure-Python package you name. No network access from inside the interpreter, so a script that calls out will fail.

download: ~12 MB on first run, then cached

Source

import sys, platform, json, collections, itertools, hashlib
from datetime import datetime, timezone

print(f"Python {sys.version.split()[0]} on {platform.machine()}")
print(f"UTC now: {datetime.now(timezone.utc):%Y-%m-%d %H:%M:%S}")

# The stdlib is all here, not a subset.
counts = collections.Counter("the quick brown fox jumps over the lazy dog".split())
print("word counts:", json.dumps(counts.most_common(3)))

pairs = list(itertools.combinations("abcd", 2))
print(f"{len(pairs)} pairs:", pairs)

print("sha256:", hashlib.sha256(b"particle wave").hexdigest()[:32], "...")

# A bare expression at the end is echoed, the way a REPL would.
sum(x * x for x in range(1, 11))

Shows what is actually available: the full stdlib, at the real version.

stdin

Ctrl Enter to run · Tab indents the selection

Output

Nothing yet. Run the sample above.

About this tool

01 · Pyodide WebAssembly

Full CPython runtime compiled to WebAssembly running directly in your browser tab. Standard library and micropip wheel installation supported.

Zero Backend: No external compute servers, accounts, or queues.

02 · Web Worker Isolation

Code executes inside an isolated background Web Worker thread. Infinite loops (while True: pass) can be safely terminated without locking the UI.

Safety: Automatic 60-second timeout ceiling + instant Stop interrupt.

03 · Package Registry

Preconfigured environments load standard packages (NumPy, SciPy, pandas) and custom project wheels dynamically from PyPI.

Live Artifacts: Real execution pipelines in a single click.

Operational Limits & Assumptions

· No Raw TCP Sockets: In-browser Python cannot open raw OS sockets; requests and raw HTTP libraries fail without wasm shims.
· C/Rust Extensions: Native binary modules require precompiled WebAssembly wheels published by maintainers.
· Execution Overhead: Wasm execution is ~2x to 5x slower than native host Python. Useful for verification, not micro-benchmarks.
· Worker Reset: Pressing Stop terminates the worker instance, clearing in-memory variables for a fresh restart.

Found a case it gets wrong? Tell me about it .