scalene
Scalene: A high-resolution, low-overhead CPU, GPU, and memory profiler for Python with AI-powered optimization suggestions
Decision gist · record as of 2026-08-14
Yes. Scalene is actively maintained, well-established (13k+ GitHub stars, top 15k PyPI packages), carries no known vulnerabilities, and solves a real problem—low-overhead, detailed profiling with AI-assisted optimization. The 9 runtime dependencies are standard and widely available. Install friction is moderate but manageable. Recommended for developers who need to understand and optimize Python performance without high profiling overhead.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Requires Python 3.8 or later (excluding 3.11.0).
- GPU profiling requires NVIDIA drivers and nvidia-ml-py.
- AI optimization suggestions require credentials for OpenAI, Azure, Amazon Bedrock, or a local Ollama instance.
License · maintenance · safety
permissive license (permissive) — Licensed under Apache License (permissive), allowing commercial and private use with minimal restrictions. No license-based constraints on adoption.
last release 2026-05-12 (94 days) · last repo commit 2026-08-01 · 13,494 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 396,130 downloads/mo, #6,972 on PyPI
Alternatives
Verify before relying
pip install scalene
# Profile a script
scalene run your_prog.py
# Or programmatically:
from scalene import scalene_profiler
scalene_profiler.start()
# your code
scalene_profiler.stop()- Whether AI optimization suggestions are available without external API credentials or if local-only operation is feasible
- Performance overhead claim of 'typically no more than 10-20%' and accuracy comparison methodology
- Compatibility with async/await code and multiprocessing workloads
What it is and what it does
Scalene is a sampling-based profiler that measures CPU, GPU, and memory usage in Python programs at the line and function level. Unlike traditional profilers that instrument code or rely on Python's tracing hooks, Scalene uses sampling to keep overhead low while delivering detailed per-line breakdowns of where execution time and memory are consumed. It runs on macOS, Linux, and Windows, supports modern Python versions (3.8–3.14), and integrates with Jupyter, IPython, and VS Code.
The package's distinguishing feature is AI-powered optimization suggestions: after profiling, users can click on specific lines or code regions to request optimization proposals from GPT-4, Azure, Amazon Bedrock, or local models via Ollama. Results are displayed in an interactive web-based GUI (or CLI) that works fully offline. Configuration is flexible—via command-line flags, YAML files, or programmatic decorators—making it suitable for one-off profiling runs and integrated performance monitoring in development workflows.
Use it for
- Identify CPU and memory bottlenecks in data science pipelines and get AI-suggested rewrites before optimization effort
- Profile GPU-accelerated code to see where computation time is spent and detect memory leaks on device
- Integrate profiling into CI/CD to catch performance regressions at the line level across Python versions
- Debug slow web services or batch jobs by profiling specific functions with the @profile decorator
- Generate shareable performance reports as self-contained HTML files for code review and documentation
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes.
Scalene is actively maintained, well-established (13k+ GitHub stars, top 15k PyPI packages), carries no known vulnerabilities, and solves a real problem—low-overhead, detailed profiling with AI-assisted optimization. The 9 runtime dependencies are standard and widely available. Install friction is moderate but manageable. Recommended for developers who need to understand and optimize Python performance without high profiling overhead.
Install
scalene on PyPI
Before you install
Medium install friction due to 9 runtime dependencies including numpy, pydantic, and nvidia-ml-py. Package is actively maintained with recent releases and strong community engagement (13494 GitHub stars). Precompiled wheels available for Python 3.8–3.14 on macOS, Linux, and Windows reduce build complexity.
Requires Python 3.8 or later (excluding 3.11.0). GPU profiling requires NVIDIA drivers and nvidia-ml-py. AI optimization suggestions require credentials for OpenAI, Azure, Amazon Bedrock, or a local Ollama instance.
License in practice
Licensed under Apache License (permissive), allowing commercial and private use with minimal restrictions. No license-based constraints on adoption.
Quickstart
pip install scalene
# Profile a script
scalene run your_prog.py
# Or programmatically:
from scalene import scalene_profiler
scalene_profiler.start()
# your code
scalene_profiler.stop()Verify before relying
- Whether AI optimization suggestions are available without external API credentials or if local-only operation is feasible
- Performance overhead claim of 'typically no more than 10-20%' and accuracy comparison methodology
- Compatibility with async/await code and multiprocessing workloads
Package facts
| License | permissive license permissive |
| Python support | Supports the current Python release !=3.11.0,>=3.8 |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | 9 packagesrichcloudpicklenvidia-ml-pyJinja2psutilnumpyastunparsepydanticpyyaml |
| Maintenance | Actively maintained 94 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 396,130 / month, #6,972 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 5 - Production/StableFramework :: IPythonFramework :: JupyterIntended Audience :: DevelopersIntended Audience :: Science/ResearchLicense :: OSI Approved :: Apache Software LicenseOperating System :: MacOS :: MacOS XOperating System :: Microsoft :: WindowsOperating System :: POSIX :: LinuxProgramming Language :: PythonProgramming Language :: Python :: 3Programming Language :: Python :: 3 :: OnlyProgramming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14Programming Language :: Python :: 3.8Programming Language :: Python :: 3.9Topic :: Software DevelopmentTopic :: Software Development :: Debuggers |
Evidence: scalene-2.3.0-cp310-cp310-macosx_15_0_universal2.whl; scalene-2.3.0-cp310-cp310-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; scalene-2.3.0-cp310-cp310-win_amd64.whl; scalene-2.3.0-cp311-cp311-macosx_15_0_universal2.whl; scalene-2.3.0-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; scalene-2.3.0-cp312-cp312-macosx_15_0_universal2.whl; scalene-2.3.0-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; scalene-2.3.0-cp312-cp312-win_amd64.whl; scalene-2.3.0-cp313-cp313-macosx_15_0_universal2.whl; scalene-2.3.0-cp313-cp313-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; scalene-2.3.0-cp313-cp313-win_amd64.whl; scalene-2.3.0-cp314-cp314-macosx_15_0_universal2.whl; scalene-2.3.0-cp314-cp314-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; scalene-2.3.0-cp314-cp314-win_amd64.whl; scalene-2.3.0-cp39-cp39-macosx_15_0_universal2.whl; scalene-2.3.0-cp39-cp39-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl; scalene-2.3.0-cp39-cp39-win_amd64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “python cpu gpu memory profiler”
- scaleneScalene profiles Python code to measure CPU, GPU, and memory…
- google-cloud-profilerCollects and uploads CPU, memory, and wall-time profiles from Python…
- austin-distAustin is a frame stack sampler for CPython that profiles running…
Give your agent the search over MCP, or paste the wish link into any chat.
More Software Development packages
Provides backported and experimental type hints for Python 3.9+, allowing use of newer typing features on older Python versions and enabling early experimentation with type system PEPs before they enter the standard library.
NumPy provides an N-dimensional array object and a comprehensive suite of mathematical, linear algebra, Fourier transform, and random number functions for scientific computing in Python.
FastAPI is a Python web framework for building REST APIs using type hints, with automatic request validation, serialization, and interactive API documentation.
Provides a way to document function parameters, class attributes, return types, and variables inline using Python's `Annotated` type hint syntax instead of traditional docstrings.
Typer builds command-line applications from Python functions using type hints, automatically generating help text, argument parsing, and shell completion.
Install it if you are building CLIs in Python.
Distlib provides low-level packaging utilities for building, distributing, and managing Python software—including metadata handling, version specifiers, wheel support, script installation, and dependency resolution.
See also torch-tb-profiler · line-profiler · google-cloud-profiler · yappi · memory-profiler · austin-dist · nvidia-nvtx · tuna · pyinstrument · Pympler