Skip to content

OS Development on MacOS

Most users can simply use a pre-built binary from the releases page and install it manually or using the web installer.

But if you want to modify things under the hood, you're in the right place!

Get the prerequisites

Make sure you have everything installed to compile code:

xcode-select --install
brew install pkg-config libffi ninja make SDL2

Download the code

Clone the repositories:

git clone --recurse-submodules --depth 1 --shallow-submodules https://github.com/MicroPythonOS/MicroPythonOS.git
cd MicroPythonOS/

That will take a while, because it recursively clones MicroPython, LVGL, ESP-IDF and all their dependencies.

Optional: updating the code

If you already have an old clone and you want to update it, the easiest is to just delete it and re-clone.

But if you need to save on bandwidth and time, you can instead do the following, which will throw away all local modifications:

cd MicroPythonOS/
git submodule foreach --recursive 'git clean -f; git checkout .'
git pull --depth 1
git submodule update --init --depth 1

Compile the code

Use the build_mpos.sh script for convenience.

Usage:

./scripts/build_mpos.sh <target system>

Target systems: esp32, esp32s3, unix (= Linux), macOS and web (= WebAssembly)

Examples:

./scripts/build_mpos.sh esp32
./scripts/build_mpos.sh esp32s3
./scripts/build_mpos.sh unix
./scripts/build_mpos.sh macOS

The resulting build file will be in lvgl_micropython/build/, for example:

  • lvgl_micropython/build/lvgl_micropy_unix
  • lvgl_micropython/build/lvgl_micropy_macOS
  • lvgl_micropython/build/lvgl_micropy_ESP32_GENERIC_S3-SPIRAM_OCT-16.bin

Running on desktop

MicroPythonOS runs on Linux, macOS and Raspberry Pi desktops. The desktop build is a single executable that contains the whole OS, so you usually do not need to compile anything.

Pick the level that matches what you want to do.

Level 1: Just want to run it?

Grab a pre-built binary and launch it.

  1. Download the binary for your platform from the releases page:
  2. Linux, Raspberry Pi, WSL2 on Windows: lvgl_micropy_unix
  3. macOS (Apple Silicon or Intel): lvgl_micropy_macOS
  4. Make it executable:
chmod +x lvgl_micropy_unix
  1. Run it. From a terminal:
./lvgl_micropy_unix

Or double-click it in your file manager. You may want to rename it to MicroPythonOS first.

Level 2: Want to develop an app but use a prebuilt OS?

You can use a downloaded OS binary with a full source checkout so you can edit apps and see changes immediately.

  1. Clone the repository:
git clone --recurse-submodules https://github.com/MicroPythonOS/MicroPythonOS.git
cd MicroPythonOS
  1. Download the binary for your platform from the releases page. Rename it and put it where the runner script expects it:
  2. Linux / Raspberry Pi / WSL2 → lvgl_micropython/build/lvgl_micropy_unix
  3. macOS → lvgl_micropython/build/lvgl_micropy_macOS
mkdir -p lvgl_micropython/build
cp /path/to/downloaded/binary lvgl_micropython/build/lvgl_micropy_unix
chmod +x lvgl_micropython/build/lvgl_micropy_unix
  1. Run it with the helper script:
./scripts/run_desktop.sh

scripts/run_desktop.sh launches the OS using the Python files in internal_filesystem/ directly, so any edit you make there appears the next time you restart the app. No rebuild is needed.

Try it:

  1. Edit internal_filesystem/builtin/apps/com.micropythonos.about/assets/about.py
  2. Run ./scripts/run_desktop.sh
  3. Open the About app
  4. See your change immediately

Once your app works on desktop, deploy it to a physical device with Installing on ESP32.

Level 3: Want to build the OS yourself?

If you need to change the OS itself, add C extensions, modify MicroPython/LVGL bindings, or run the very latest code, build from source. The binary will already land in lvgl_micropython/build/lvgl_micropy_XXX where XXX is unix or macOS.

./scripts/build_mpos.sh unix     # Linux, Raspberry Pi, WSL2
./scripts/build_mpos.sh macOS    # macOS

See Compiling for the full instructions, including cloning and dependencies.

Known issues and fixes

Linux: "No such file or directory" when running the binary

Make sure it is executable:

chmod +x lvgl_micropy_unix

macOS: missing libffi.8.dylib

Install libffi:

brew install libffi

macOS: "cannot be opened because the developer cannot be verified"

The prebuilt binary is not signed. Open System Settings → Privacy & Security and click Allow Anyway next to the blocked item, then run it again.

Allow Anyway on MacOS

Windows

Native Windows builds are not supported. Users report that the Linux desktop binary works under WSL2 on Windows 11. Alternatively, you can try the web port, which runs a desktop build in the browser.

Deploying to hardware

Once your app works on desktop, install it on a supported ESP32 device.

The easiest way to install on the ESP32 is using the webinstaller, of course.

But if you need to install a version that's not available there, or you built your own, then you can manually install it on an ESP32 device.

  1. Get the firmware

    • Download a release binary (e.g., MicroPythonOS_esp32_0.5.0.bin)
    • Or build your own on MacOS or Linux

  2. Put the ESP32 in Bootloader Mode

    If you're already in MicroPythonOS: go to Settings - Restart to Bootloader - Bootloader - Save.

    Otherwise, physically keep the "BOOT" (sometimes labeled "START") button pressed while powering up the board. This is explained in more detail at the webinstaller

  3. Flash the firmware

    ~/.espressif/python_env/idf5.2_py3.9_env/bin/python -m esptool --chip esp32s3 write_flash 0 firmware_file.bin

    Add the --erase-all option if you want to erase the entire flash memory, so that no old files or apps will remain.

    There's also a convenient ./scripts/flash_over_usb.sh script that will attempt to flash the latest firmware that you compiled yourself.

  4. Access the MicroPython REPL shell

    After reset, the REPL shell should be available on the serial line.

    Any serial client will do, but it's convenient to use the mpremote.py tool that's shipped with lvgl_micropython:

    ./lvgl_micropython/lib/micropython/tools/mpremote/mpremote.py

  5. Populate the filesystem (only for development)

    In development, you probably want to override the "frozen" libraries and apps that are compiled in, and replace them with source files, which you can edit.

    This makes MicroPythonOS startup a lot slower, as the Python scripts have to be compiled at runtime instead of at build time. But once MicroPythonOS and the code you're testing has loaded, the speed will be normal again.

    There's a convenient script that will do this for you.

    Usage:

    ./scripts/install.sh


    ./scripts/install.sh com.micropythonos.about # to install one single app

    On MacOS, the install.sh script needs: brew install --cask serial

    If you need to frequently update a small number of files, you can also update them manually, for example:

    ./lvgl_micropython/lib/micropython/tools/mpremote/mpremote.py cp internal_filesystem/lib/mpos/device_info.py :/lib/mpos

Notes

Back to top