Python Commands Cheatsheet (CLI, Interpreter, pip, venv)
This covers the python executable itself, package management, and environment tooling — for language syntax, see the separate Python language cheatsheet.
Running Python
python script.py # run a script
python3 script.py # explicit Python 3 (some systems still alias python -> Python 2)
python - # read script from stdin
python -c "print('hello')" # run an inline command
python -m module_name # run a module as a script (uses sys.path properly)
python -m http.server 8000 # quick local web server
python -m json.tool file.json # pretty-print JSON from the CLI
python -m venv myenv # create a virtual environment
python -m pip install package # run pip via the module (avoids PATH issues)
python # interactive REPL
python -i script.py # run script, then drop into interactive mode with its namespace
python -O script.py # optimized mode (strips assert statements, __debug__ = False)
python -OO script.py # also strips docstrings
Interpreter Flags
python -V # version
python --version
python -h # help
python -u script.py # unbuffered stdout/stderr (useful for real-time log streaming, e.g. Docker/Cloud Run)
python -B script.py # don't write .pyc files
python -E script.py # ignore PYTHON* environment variables
python -s script.py # don't add user site-packages to sys.path
python -W ignore script.py # suppress warnings
python -W error script.py # turn warnings into errors
python -X dev script.py # enable development mode (extra runtime checks)
python -X importtime script.py # profile import time — great for diagnosing slow startup
python -m pdb script.py # run under the debugger from the start
python -m cProfile script.py # profile the script
python -m cProfile -o out.prof script.py # save profile output for later analysis (e.g. snakeviz)
python -m timeit "sum(range(100))" # time a single expression from the CLI
python -m py_compile script.py # compile to bytecode without running
python -m trace --trace script.py # trace every executed line
Getting Info
python -c "import sys; print(sys.path)" # module search path
python -c "import sys; print(sys.executable)" # which interpreter is actually running
python -c "import sys; print(sys.version)"
python -c "import platform; print(platform.python_version())"
python -c "help('modules')" # list all installed modules (slow)
python -m site # site-packages locations, sys.path
which python # (Linux/Mac) which interpreter PATH resolves to
where python # (Windows) same
python -c "import os; print(os.getcwd())"
pip — Package Management
pip install package
pip install package==1.2.3
pip install "package>=1.2,<2.0"
pip install --upgrade package
pip install --upgrade pip
pip uninstall package
pip install -r requirements.txt # see the dedicated requirements.txt cheatsheet for format details
pip install -e . # editable install of the current project (from pyproject.toml/setup.py)
pip install -e ".[dev]" # editable install with an optional extras group
pip list # installed packages
pip list --outdated # what has newer versions available
pip show package # detailed metadata about an installed package
pip freeze # exact installed versions, pip-freeze format
pip check # verify dependency compatibility across installed packages
pip download package -d ./pkgs # download without installing
pip install --no-index --find-links=./pkgs package # install from local files, no network
pip cache list
pip cache purge
pip config list
pip install --break-system-packages package # needed on some modern Debian/Ubuntu externally-managed environments
pip install --index-url https://pypi.example.com/simple package # custom package index
pip install --proxy http://proxy.example.com:8080 package # through a proxy
Virtual Environments
python -m venv myenv # create
source myenv/bin/activate # activate (Linux/Mac)
myenv\Scripts\activate # activate (Windows cmd)
myenv\Scripts\Activate.ps1 # activate (Windows PowerShell)
deactivate # exit the venv
python -m venv myenv --system-site-packages # inherit globally installed packages
python -m venv myenv --clear # wipe and recreate
python -m venv myenv --upgrade # upgrade venv's bundled pip/setuptools to match interpreter
# Check whether you're in a venv
python -c "import sys; print(sys.prefix != sys.base_prefix)"
echo $VIRTUAL_ENV # set while a venv is activated
uv (modern, fast alternative to venv/pip)
pip install uv
uv venv # create a virtual environment (defaults to .venv)
uv venv --python 3.12 # specify Python version
uv pip install package # much faster than pip
uv pip install -r requirements.txt
uv pip freeze
uv pip sync requirements.txt # install exactly what's pinned, removing extras
uv run script.py # run a script inside the project's environment automatically
uv add package # add + install a dependency, updates pyproject.toml
uv lock # generate a lock file
pipx (isolated CLI tool installs)
pip install pipx
pipx install black # installs `black` into its own isolated venv, exposes the CLI globally
pipx list
pipx upgrade black
pipx uninstall black
pipx run black --check . # run a tool once without installing it permanently
Package/Project Management (pyproject.toml era)
pip install build
python -m build # build sdist + wheel from pyproject.toml
pip install twine
twine upload dist/* # publish to PyPI
twine upload --repository testpypi dist/* # publish to TestPyPI
# Poetry (alternative full-featured tool)
poetry init
poetry add requests
poetry add --group dev pytest
poetry install
poetry run python script.py
poetry build
poetry publish
poetry shell # activate the managed virtual environment
# pipenv (another alternative)
pipenv install requests
pipenv install --dev pytest
pipenv shell
pipenv run python script.py
pipenv lock
Version Management (multiple Python versions)
# pyenv (Linux/Mac)
pyenv install 3.12.3
pyenv versions
pyenv global 3.12.3 # set default globally
pyenv local 3.11.8 # set version for current directory (writes .python-version)
pyenv shell 3.10.13 # set version for current shell session only
pyenv which python
# On Windows: py launcher (bundled with the official installer)
py -3.12 script.py
py -3.11 -m venv myenv
py --list # list installed Python versions
py -0 # same, shorter alias
Linting, Formatting, Type Checking (CLI usage)
pip install black ruff mypy isort
black . # auto-format code
black --check . # check without modifying (CI-friendly)
black --diff . # show what would change
ruff check . # lint (fast, replaces flake8 + many plugins)
ruff check --fix . # auto-fix issues
ruff format . # ruff also has a formatter (Black-compatible)
isort . # sort imports
mypy script.py # static type checking
mypy --strict script.py # stricter checks
pytest # run tests
pytest -v # verbose
pytest -k "test_name" # run tests matching a keyword
pytest --cov=mypackage # coverage report (needs pytest-cov)
pytest -x # stop at first failure
pytest --lf # rerun only last-failed tests
Common Environment Variables
PYTHONPATH=/extra/module/dir # additional module search paths
PYTHONDONTWRITEBYTECODE=1 # skip writing .pyc files
PYTHONUNBUFFERED=1 # unbuffered stdout/stderr (common in Docker/Cloud Run configs)
PYTHONWARNINGS=ignore # suppress warnings
PYTHONHASHSEED=0 # deterministic hash seed (useful for reproducible tests)
PIP_NO_CACHE_DIR=1 # disable pip's cache (common in Docker builds to shrink layers)
PIP_INDEX_URL=https://pypi.example.com/simple # custom default package index
VIRTUAL_ENV # set automatically when a venv is activated
Common Gotchas
pythonvspython3— on many Linux distros and macOS,pythonmay not exist or may point to Python 2; usepython3explicitly, or check withpython --versionfirst.pip installwithout an active virtual environment installs globally (or fails on “externally managed” systems, requiring--break-system-packagesor a venv) — always activate a venv first for project work.python -m pipis more reliable than barepipwhen multiple Python installations exist, since it guarantees you’re using the pip tied to that specific interpreter.- Mixing
pip install --userand virtual environments can cause confusing precedence issues — avoid--userinstalls when working inside a venv. pip freezecaptures the exact environment, including transitive dependencies — good for lockfiles, not ideal as a hand-maintained source of truth (see therequirements.txtcheatsheet for the pip-tools workflow).