BenchCode · iPad and iPhone
BenchCode teaches programming and scientific computing through short lessons you run yourself. Each lesson is a small real program with notes beside it: read a step, tap Run, change a number, run it again and watch the result change.
Twelve guided lessons in six tracks: Python basics, data and charts, probability, maths and LaTeX, machine learning, and C programming. Every step says what to try next, your progress is saved, and each lesson's code is yours to edit or reset.
A program's result appears beside its code: printed text, a table, a chart or a typeset PDF.
Ask about any line of your program, ask for a hint, or have your answer checked. The tutor explains; it doesn't take over. Code it suggests goes into the editor and runs only when you tap Run.
It uses Apple Intelligence where available, or an open model you download once and keep on your device.
Work through exercises cell by cell with text, maths, tables and charts in one notebook. The libraries used in university courses are built in: NumPy, SciPy, pandas, scikit-learn, statsmodels, PyTorch, SymPy, matplotlib and Plotly, with LaTeX for writing up reports.
No account, no subscription and no ads. Your code and results stay on your device, so BenchCode works in class, in a lab or anywhere without a signal. See the BenchCode privacy policy.
Questions, a lesson that doesn't run, or an idea for a new one: email euler.yu@gmail.com and say which device and iPadOS or iOS version you use.
If the tutor answers slowly, open Settings › Tutor model: it shows how much of your device's memory the model uses, and a smaller model, or unloading it when you're done, keeps BenchCode responsive.