Find it slow.
Make it fast.
ACENLY benchmarks your Python functions and rewrites the slow ones algorithmically — O(n²) to O(n), verified correct, all on your machine. Use it from the command line or directly inside VS Code.
Benchmark tool is open source and free. Optimizer is €49 one-time.
Both include the full optimizer, benchmarks, and local web dashboard — one runs standalone, one lives inside VS Code. The benchmark tool alone is free and open source.
ACENLY
Run the optimizer from the command line. A local web dashboard opens in your browser — optimizer UI, run history, speedup charts, and diff viewer all in one place.
python main.py myfile.py to startACENLY for VS Code
Everything in the standalone, but built right into your editor. Right-click any function to optimize it, run benchmarks from a VS Code panel, and see the diff — without leaving the editor.
Every step runs locally. Your code never leaves your machine.
Benchmark
ACENLY profiles your functions across realistic input shapes — small, medium, large — and shows you exactly which ones are slow and why.
Optimize
The evolution engine generates dozens of algorithmically different rewrites — different data structures, loop patterns, stdlib tricks — and benchmarks every one.
Verify
Before accepting any result, ACENLY runs full equivalence checks: same outputs across hundreds of cases, including edge cases and adversarial inputs.
A local web dashboard and VS Code extension — not a command you run and forget.
Optimizer UI
Pick a function, pick a mode (Safe → Max), and watch candidates generate and race in real time.
Results viewer
Winner shown with speedup badge, side-by-side diff, full benchmark log, and correctness proof.
Dashboard
Every run, every function, every speedup — tracked over time so you know exactly what changed.
100% local
Runs entirely on your machine. No data sent anywhere. Use local Ollama models for zero API cost.
Algorithmic rewrites
Not cosmetic changes. Real algorithmic improvements — faster data structures, better loop patterns, stdlib upgrades.
Correctness-verified
Equivalence-checked on your test cases plus adversarial edge cases — before any result is accepted.