Getting started

Installation

Install LineScope in your script or notebook's environment. Python 3.11–3.15 is supported.

Latest release

uv pip install -U linescope

Latest source

For unreleased changes from main:

uv pip install -U git+https://github.com/tvdboom/linescope.git@main

Optional dependencies

Extra	Adds
scalene	CPU sampling and supported GPU metrics on Python 3.11–3.14
notebook	IPython cell magics and inline reports
full	Both extras

uv pip install -U "linescope[full]"

Scalene is omitted on Python 3.15. See dependencies for platform support.

Usage

From Python

Wrap the code to measure in profile:

from linescope import profile

with profile(memory=True):
    values = list(range(100_000))
    total = sum(value * value for value in values)

The report opens when the block ends. Trace is the default backend; memory=True adds process RAM and retained Python allocation measurements.

To save without opening a browser:

from linescope import profile

with profile(display="none") as session:
    total = sum(value * value for value in range(100_000))

session.save("linescope.html")

From the CLI

linescope your_script.py
linescope -m your_package.your_module

The report opens when the program exits. Put profiler options before the target; arguments after it go to your program:

linescope --memory --output report.html your_script.py --rows 100000

Use --display none to suppress the browser, or --backend scalene to select sampling. See the CLI reference for all options.

From a notebook

Install the notebook extra, then load the extension:

%load_ext linescope

Profile a cell:

%%profile --display cell-summary

total = sum(value * value for value in range(100_000))

For several cells, start a session, run your cells, then stop:

from linescope import profile

session = profile.start(display="end")

profile.stop()

See notebooks for display modes and Databricks.