How to Run Python on iPhone and iPad

Juno is a zero-setup Python IDE, Jupyter environment and developer toolbox in a native app for iPad and iPhone. It runs Python 3.13 locally on your device, offline, and supports Python scripts (.py), multi-file projects and Jupyter notebooks (.ipynb) in one iOS experience.

Quick answer

If you are asking “how do I run Python on iPhone or iPad?”, the short version is: install Juno from the App Store, open a script or notebook, and tap Run. There is nothing to set up—no server, no account, no terminal—and your code runs on the device itself, so it works on a plane as well as at a desk.

Juno gives you:

  • Python scripts, projects and Jupyter notebooks in one app
  • A built-in Python 3.13 interpreter that runs locally and offline
  • A full development workspace with code completion, a console with terminal emulation and an Output sidebar
  • A notebook editor designed for iPad and iPhone, with touch, keyboard and trackpad support
  • More than 150 bundled packages, from NumPy and pandas to pygame, Keras and llama.cpp
  • A package manager for the vast majority of pure-Python packages on PyPI
  • iOS automation: run Python from Shortcuts and Siri, and reach the clipboard, contacts, photos, location and sensors from your code
  • A standalone Python REPL, Object Inspector and an AI coding assistant

Editing is free. Running code is free to try, and continued execution unlocks with a one-time Juno Pro purchase—not a subscription.

What is free

  • Navigate and manage files and folders, with quick document and media previews
  • Browse, view and edit Jupyter notebooks with beautifully rendered previews
  • View and edit Python, JSON, Markdown, XML and other text-based formats in one code-and-text editor
  • Add quick-access bookmarks for file-system locations outside the app sandbox, or open documents in place through Apple’s Files app
  • Try running code in scripts, notebooks, projects and the REPL

Run Python scripts and projects

Juno lets you work with .py files in a full Python workspace: intelligent code completion, console output with terminal emulation, and rich output handling for plots, images and local web apps in a built-in browser. Open a folder as a project to get a file tree, multiple editor tabs, a project-aware coding assistant and per-project run configuration. When a script finishes, you can drop into a REPL that keeps its definitions, or start a standalone REPL session from Home.

Read more: Working with Python scripts and projects and the Juno Python API reference.

Automate iPhone and iPad with Python

Juno is also an iOS automation tool: Python scripts can be triggered from Shortcuts and Siri, and they can talk to the device they run on.

Shortcuts and Siri. Juno’s Shortcuts actions run Python code or a .py script, pass in arguments and files, capture the output and hand it to the next action. A script can become a Siri command, a Shortcuts automation that fires on a schedule or an event, or one step in a longer workflow built from Apple’s own actions, and long-running scripts can keep working in the background.

Device APIs from Python. The built-in juno modules give Python access to the clipboard, Keychain, device information, location, motion sensors, contacts, reminders, photos, sound, notifications, sharing and Bluetooth. Scripts can present native alerts, pickers and full-screen image, video and web previews, and an Objective-C bridge opens up UIKit, Foundation and other system frameworks when you need more.

Files. Juno edits documents in place through Apple’s Files app, and bookmarks keep folders outside the app sandbox, including iCloud Drive and other providers, one tap away, so scripts can process the files you already have on the device.

Read more: the Juno Python API reference.

Edit and run Jupyter notebooks

Jupyter notebooks combine live code, Markdown, plots, images and equations in .ipynb files. Juno provides a modern notebook editor on iOS and iPadOS for creating, editing and executing notebook cells, with rich outputs preserved and a kernel that runs on your device. Notebooks also open as editable tabs inside projects, each with its own kernel started in the project’s folder, so several can run side by side.

Read more: Working with notebooks.

Make games with pygame and turtle

pygame is bundled with Juno, so you can build games in Python with touch controls, sound and smooth full-screen graphics, and play them on the same device you wrote them on. Eight complete games ship in the Examples folder—Breakout, Snake, 2048, Flappy, Birdle, Match-3, Solitaire and Block Land—along with three retro emulators for the NES, Game Boy and CHIP-8, each written in Python with pygame. Play any of them straight away, then open the source to see how it works. pymunk is included for 2D physics.

Turtle graphics work out of the box too: import turtle needs no setup on iPad or iPhone, and a Turtle starter project collects twenty-five short scripts, one idea per file.

Run local LLMs and machine learning on device

Juno runs language models, image generation and neural-network training on your iPhone or iPad, from your own Python code, with no server involved:

  • llama-cpp-python, the Python bindings for llama.cpp, runs GGUF language models with Metal GPU acceleration and streams replies token by token. Any .gguf file works, including your own.
  • MLX, Apple’s array framework for Apple silicon, comes with mlx-lm, which runs the ready-converted models published by the mlx-community on Hugging Face and fine-tunes them with LoRA adapters on device.
  • Keras 3 runs on an MLX backend, so the same Keras code you would write on a laptop trains on your iPad.
  • Stable Diffusion generates images on MLX inside a bundled example project.
  • huggingface_hub, transformers, tokenizers and safetensors are built in.

The place to start is the AI & ML section of Home, with five ready-to-run projects: Run LLMs Locally, Stable Diffusion, Keras on MLX, RAG with Embeddings and LoRA Fine-Tuning. Models download the first time a project needs them, after it tells you the size and asks, and work offline after that. The default models fit an iPhone or iPad with 4 GB of memory.

PyTorch, TensorFlow and JAX are not available on iOS; the FAQ explains why and what to use instead.

Bundled packages

Juno ships with more than 150 Python packages pre-installed, so there is nothing to download before you import:

  • Scientific computing and machine learning: NumPy, SciPy, Scikit-learn, XGBoost, LightGBM, CVXOPT
  • On-device AI: llama-cpp-python, MLX, mlx-lm, Keras, huggingface_hub, transformers, tokenizers, safetensors
  • Data analysis: pandas, Polars, DuckDB, Statsmodels, PyArrow, h5py
  • Plotting and image processing: Matplotlib, Seaborn, OpenCV, Pillow, scikit-image
  • Games and graphics: pygame, pymunk, turtle
  • Geospatial: GDAL, GeoPandas, Shapely, Fiona
  • Web and APIs: requests, openai, tiktoken
  • Bioinformatics: Biopython
  • And more: LXML, Cryptography, PyCryptodome, Pydantic, Regex, pypdfium2, z3-solver, msgpack, zstandard

The full list lives in Juno’s package manager, which also installs the vast majority of pure-Python packages from PyPI, resolves their dependencies automatically and can install the supported packages from a project’s requirements.txt, Pipfile or environment.yml.

Juno vs other ways to run Python on iOS

There are several good iOS tools, including Pythonista, Pyto and a-Shell, and browser-based services such as Google Colab. The right choice depends on your workflow.

  • If you want local Python, Jupyter notebooks, iOS automation, pygame and on-device machine learning together in one iOS app, Juno is a strong fit.
  • If you want to connect to an existing remote Jupyter server, use Juno Connect, a separate app by the same developer.
  • If you are comparing alternatives, check each app’s latest App Store page for current features and pricing.

Juno is designed as a first-class iOS app, not a remote terminal or thin client. That makes it a strong option for offline coding on iPad and iPhone, classroom use, and mobile workflows that still need real developer tooling: keyboard and trackpad support, multitasking, in-place editing through Files, and background execution for long-running scripts.

Important limitations

  • Some PyPI packages are not compatible with iOS platform restrictions
  • Packages that rely on native extensions cannot generally be installed at runtime unless they are bundled with Juno
  • PyTorch, TensorFlow and JAX are not available; use the bundled MLX, Keras and llama-cpp-python instead
  • Conda environments and additional Jupyter kernels are not supported
  • Juno requires iOS or iPadOS 17 or later

Pricing

Juno is free to download, and editing notebooks and files is free. Running code is free to try in notebooks, scripts, projects and the REPL; continued execution requires a one-time Juno Pro purchase, not a subscription. Please refer to the official App Store listing for the latest price and feature availability: