py-runner

Persistent Python code runner: string in, JSON out. A lightweight, dependency-free, ipykernel-like execution library — no Jupyter server, no ZMQ, no required third-party packages.

📖 Full documentation: https://rightfix.github.io/py-runner-docs/

Persistent Python code runner: string in, JSON out. A lightweight, dependency-free, ipykernel-like execution library — no Jupyter server, no ZMQ, no required third-party packages.

from py_runner import Kernel

k = Kernel()
k.execute("x = 40 + 2")
print(k.execute("x + 1"))
# {"output": null,
#  "display": [{"type": "text/plain", "data": "43"}],
#  "error": null, "execution_count": 2, "execution_time": 0.0001}

Install

pip install py-runner-kernel

Requires Python >= 3.10. Zero runtime dependencies.

Quickstart

from py_runner import Kernel, execute, execute_json

# Module-level shared kernel
execute("import math")
print(execute_json("math.sqrt(16)"))

# Isolated session kernel (one per user/request)
k = Kernel(on_input=lambda prompt: "bob", max_history=100)
r = k.execute('name = input("who? ")\nname.upper()')
assert r["display"][0]["data"] == "'BOB'"

Each result is a JSON-serializable dict:

Features

• Persistent namespace with Jupyter-style _, __, ___ and bounded Out[N] history
• Last-expression display — DataFrames as HTML, arrays/tensors summarized, matplotlib figures as PNG, trailing ; suppresses output
• Magics: %time, %timeit, %who/%whos, %run, %pip, !shell, %reset
• Long runs: timeout=, cooperative interrupt(), on_stream chunks, on_heartbeat for silent fit() loops, tqdm \r collapsing, logging/warnings capture
• Sandboxing: allow_shell=False, allow_pip=False, set_address_space_limit(mb) (Unix), reset(keep_imports=True)
• CLI: py-runner -c "1+1", py-runner script.py, ... | py-runner, with --timeout, --no-shell, --no-pip

CLI

py-runner -c "print(2 + 3)"
py-runner analysis.py --timeout 300 > result.json
cat cell.py | py-runner
python -m py_runner -c "1+1"

Project structure

py_runner/
├── __init__.py   # public API: Kernel, execute, execute_json, reset, main
├── kernel.py     # persistent execution kernel (threaded, timeouts, input)
├── magics.py     # %time, %pip, !shell, … (sandbox-flaggable)
├── display.py    # rich display collector (pandas/numpy/PIL/matplotlib)
├── streams.py    # thread-routed streaming stdout, tqdm \r handling
└── utils.py      # guarded imports, address-space limit helper

Development

uv sync            # create .venv
uv run python -c "from py_runner import execute; print(execute('1+1'))"
uv build           # wheel + sdist in dist/
uv publish --publish-url https://test.pypi.org/legacy/  # trial run

HTTP server

py-runner-serve --host 127.0.0.1 --port 8000 --max-sessions 100

S=$(curl -s -X POST localhost:8000/sessions | python3 -c "import sys,json; print(json.load(sys.stdin)['session_id'])")
curl -s -X POST localhost:8000/execute \
  -d "{\"session_id\":\"$S\",\"code\":\"x = 40 + 2\"}"
curl -s -X POST localhost:8000/execute \
  -d "{\"session_id\":\"$S\",\"code\":\"x + 1\"}"        # display: 43 (state kept)
curl -s -X POST localhost:8000/interrupt \
  -d "{\"session_id\":\"$S\"}"
curl -s localhost:8000/sessions                          # list live sessions
curl -s -X DELETE localhost:8000/sessions/$S             # destroy session

Sessions are server-generated UUIDs, live until deleted (no idle timeout), and isolated per ID. Kernels default to allow_shell=false, allow_pip=false; pass {"allow_shell": true} to POST /sessions to relax per session.

License

AGPL-3.0-or-later — see LICENSE. Note the network clause (§13): if you serve a modified version over a network, you must offer its source to those users.