Nifty Python Profiler

Profile Python scripts and pytest runs with cProfile: hotspot table, call counts and timings inline next to each function.

Maintained. No sign-in. No telemetry. Works in VS Code, Cursor, Windsurf, VSCodium and other editors that use Open VSX.

Features

Each profiled function's total time, share of the run, own time and calls at the end of its def line, the hot ones highlighted

The hotspot table beside the code, sorted by total time with a bar for each function's share

One-click profiling. Run the current Python file under cProfile from the editor's Run menu (next to the play button), the right-click menu or the Explorer. Pass arguments with Profile Current File with Arguments…; they're remembered per file.
Profile pytest runs. Right-click a test_*.py file to profile just that file, or profile the whole suite. Extra pytest arguments go in nifty.python-profiler.pytestArgs.
Hotspot table. Every function with its calls, own time, total time and time per call. Sort by any column, filter by name or file, hide libraries and built-ins, and click a row to jump to the code and see who called it.
Timings inline in your code. Each profiled function shows ⏱ 1.20 s (38.5%) · own 12.0 ms · 3 calls at the end of its def line. Functions over 10% of the run are highlighted, and hovering shows the details and the callers.
Works with saved profiles. Open any .prof / .pstats file (from python -m cProfile -o, Django or pytest plugins), or save a run to open later in SnakeViz or pstats.
Uses your interpreter. Picks up the interpreter selected in the Python extension (venvs, conda, uv), falls back to python on your PATH, or set one explicitly.
Nothing to install. cProfile ships with Python, so there's no package to add to your project.

Command	What it does
`Nifty Python Profiler: Profile Current File`	Run the current file under cProfile and show the results
`Nifty Python Profiler: Profile Current File with Arguments…`	Same, with command-line arguments for the script
`Nifty Python Profiler: Profile pytest Run`	Profile the current test file, or the whole pytest suite
`Nifty Python Profiler: Open Profile File (.prof)…`	Load a saved cProfile / pstats file
`Nifty Python Profiler: Show Profile Results`	Reopen the hotspot table
`Nifty Python Profiler: Save Profile As…`	Save the current run as a `.prof` file
`Nifty Python Profiler: Toggle Inline Timings`	Show or hide timings in the editor
`Nifty Python Profiler: Clear Profile`	Remove the timings from the editor

Setting	Default	What it does
`nifty.python-profiler.pythonPath`	`""`	Interpreter to use. Empty uses the Python extension's selection, then `python`/`python3`
`nifty.python-profiler.pytestArgs`	`[]`	Extra arguments for pytest runs
`nifty.python-profiler.inlineTimings`	`true`	Show timings at the end of each profiled `def` line
`nifty.python-profiler.hotThresholdPercent`	`10`	Highlight functions that take at least this share of the run
`nifty.python-profiler.openResults`	`true`	Open the hotspot table after each run
`nifty.python-profiler.showProgramOutput`	`true`	Show the program's output while profiling

Own time is time spent in the function's own code; total time includes everything it calls. Start with the biggest total time, then drill into the callee with the biggest own time.
cProfile measures function calls, so very fast functions called millions of times look slower than they are. Compare with and without profiling before optimizing them.

VS Code: search for "Nifty Python Profiler" in the Extensions view, or install from the Visual Studio Marketplace.
Cursor, Windsurf, VSCodium, Kiro, Antigravity: install from Open VSX.

This extension collects no telemetry and needs no account.

Nifty pytest: pytest Test Explorer: A pytest test explorer with saved run configurations, marker filters, real assertion diffs, rerun-failed and slowest-test reports. (Open VSX)
Nifty Code Runner: Run Code in 25+ Languages: Run the current file or selection in 25+ languages in the terminal, with input support, your own commands and no telemetry. (Open VSX)
Nifty Env: .env Files, Profiles & Secrets: Load .env files into terminals, tasks and debug sessions, switch between .env profiles, hide secret values on screen, and check .env against .env.example. (Open VSX)
Nifty Import Cost: Bundle Size of Imports: See the bundle size of every import inline, fast and cached, without maxing out your CPU. (Open VSX)
Nifty Launch: launch.json Generator & Checker: Add debug configurations for what's in your project (npm scripts, tests, Python, Django, FastAPI, Go, .NET, Rust, browsers), build compounds, and catch broken launch.json references. (Open VSX)
Nifty Monitor: CPU & Memory in the Status Bar: CPU, memory and disk usage in the status bar, including on remote SSH machines and dev containers. (Open VSX)

See all 100+ Nifty extensions and web tools at https://getnifty.dev