Table Of Contents
Kivy on Android¶
You can run Kivy applications on Android, on (more or less) any device with OpenGL ES 2.0 (Android 2.2 minimum). This is standard on modern devices; Google reports the requirement is met by 99.9% of devices.
Kivy APKs are normal Android apps that you can distribute like any other, including on stores like the Play store. They behave properly when paused or restarted, may utilise Android services and have access to most of the normal java API as described below.
Package for Android¶
The Kivy project provides all the necessary tools to package your app on Android, including building your own standalone APK that may be distributed on a market like the Play store. This is covered fully in the Create a package for Android documentation.
Debugging your application on the Android platform¶
You can view the normal output of your code (stdout, stderr), as well as the normal Kivy logs, through the Android logcat stream. This is accessed through adb, provided by the Android SDK. You may need to enable adb in your device’s developer options, then connect your device to your computer and run:
You’ll see all the logs including your stdout/stderr and Kivy logger.
If you packaged your app with Buildozer, the adb tool may not be in
$PATH and the above command may not work. You can instead run:
buildozer android logcat
to run the version installed by Buildozer, or
find the SDK tools at
You can also run and debug your application using the Kivy Launcher. If you run your application this way, you will find log files inside the “/.kivy/logs” sub-folder within your application folder.
Using Android APIs¶
Although Kivy is a Python framework, the Kivy project maintains tools to easily use the normal java APIs, for everything from vibration to sensors to sending messages through SMS or email.
For new users, we recommend using Plyer. For more advanced access or for APIs not currently wrapped, you can use Pyjnius directly. Kivy also supplies an android module for basic Android functionality.
User contributed Android code and examples are available on the Kivy wiki.
Plyer is a pythonic, platform-independent API to use features commonly found on various platforms, particularly mobile ones. The idea is that your app can call simply call a Plyer function, such as to present a notification to the user, and Plyer will take care of doing so in the right way regardless of the platform or operating system. Internally, Plyer uses Pyjnius (on Android), Pyobjus (on iOS) and some platform specific APIs on desktop platforms.
For instance, the following code would make your Android device vibrate, or raise a NotImplementedError that you can handle appropriately on other platforms such as desktops that don’t have appropriate hardware::
from plyer import vibrator vibrator.vibrate(10) # vibrate for 10 seconds
Plyer’s list of supported APIs is growing quite quickly, you can see the full list in the Plyer README.
Pyjnius is a Python module that lets you access java classes directly from Python, automatically converting arguments to the right type, and letting you easily convert the java results to Python.
Here is a simple example showing Pyjnius’ ability to access the normal Android vibration API, the same result of the plyer code above:
# 'autoclass' takes a java class and gives it a Python wrapper from jnius import autoclass # Context is a normal java class in the Android API Context = autoclass('android.content.Context') # PythonActivity is provided by the Kivy bootstrap app in python-for-android PythonActivity = autoclass('org.renpy.android.PythonActivity') # The PythonActivity stores a reference to the currently running activity # We need this to access the vibrator service activity = PythonActivity.mActivity # This is almost identical to the java code for the vibrator vibrator = activity.getSystemService(Context.VIBRATOR_SERVICE) vibrator.vibrate(10000) # The value is in milliseconds - this is 10s
This code directly follows the java API functions to call the vibrator, with Pyjnius automatically translating the api to Python code and our calls back to the equivalent java. It is much more verbose and java-like than Plyer’s version, for no benefit in this case, though Plyer does not wrap every API available to Pyjnius.
Pyjnius also has powerful abilities to implement java interfaces, which is important for wrapping some APIs, but these are not documented here - you can see Pyjnius’ own documentation.
Python-for-android includes a python module (actually cython wrapping java) to access a limited set of Android APIs. This has been largely superseded by the more flexible Pyjnius and Plyer as above, but may still occasionally be useful. The available functions are given in the python-for-android documentation.
This includes code for billing/IAP and creating/accessing Android services, which is not yet available in the other tools above.
Status of the Project and Tested Devices¶
These sections previously described the existence of Kivy’s Android build tools, with their limitations and some devices that were known to work.
The Android tools are now quite stable, and should work with practically any device; our minimum requirements are OpenGL ES 2.0 and Android 2.2. These are very common now - Kivy has even been run on an Android smartwatch!
A current technical limitation is that the Android build tools compile only ARM APKs, which will not run on Android devices with x86 processors (these are currently rare). This should be added soon.
As Kivy works fine on most devices, the list of supported phones/tablets has been retired - all Android devices are likely to work if they meet the conditions above.