You can automate Android from Python without rooting the device.
For many workflows, you do not need private app data, privileged system permissions, or a modified OS. You need to tap visible UI, type into fields, wait for screens, and capture enough evidence when something fails.
Handsets provides a small Python wrapper around a fast Android CLI so you can write those flows in normal Python.
Most LLM-driven Android automation starts by showing the model a screen.
That sounds reasonable. A human looks at the phone, decides what to tap, and taps it. Give the model the same view.
The problem is that "the same view" is expensive.
A full screenshot is expensive. A raw Android UI XML dump is also expensive, just in a quieter way. The model reads thousands of tokens of layout machinery before it reaches the handful of labels that matter:
For one step, that waste is easy to ignore. For a 50-step mobile agent trajectory, it becomes the bill.
uiautomator2 is a popular Python library for Android UI automation.
It is a good choice when you want a Python-native API around Android's UIAutomator framework. It can find elements by text, click buttons, inspect UI hierarchy, and drive common app flows.
But it is not always the best fit.
If your workflow is CLI-first, latency-sensitive, or driven by an LLM agent, a smaller Android automation tool may be easier to compose.
Handsets is one such alternative.