github.devotd.
NewAI agents

minitap-ai/mobile-use v3.3.0 — Natural-language mobile agent

minitap-ai/mobile-use is a Python agent that operates Android and iOS apps with natural-language commands and structures the information it finds.

Repositoryminitap-ai/mobile-use ↗

minitap-ai/mobile-use is a Python agent that operates Android and iOS apps with natural-language commands and structures the information it finds.

GitHub description: AI agents can now use real Android and iOS apps, just like a human.

stars
★ 3.3K
Forks
301
Language
Python
License
Apache-2.0
Latest release
v3.3.0
Read original source ↗View on GitHub ↗

minitap-ai/mobile-use is an open-source AI agent that controls Android or iOS devices through natural-language commands. The latest release shown on the repository is v3.3.0. However, the provided facts do not record what newly changed in this version, so the specific changes of v3.3.0 cannot be described from this information alone. The repository name and version are confirmed, but it is more accurate not to assert the details of the release as a change log.

This project focuses on navigating an app's user interface to carry out tasks, and on extracting information from apps and organizing it into the desired structure. In other words, it is an agent that executes tasks requested in natural language inside mobile apps, or that structures app data. The README summary states that multiple LLM providers can be configured. The implementation language is Python, and it is released under the Apache-2.0 license. The repository shows 3286 stars.

This description shows the purpose of the project and the feature scope recorded in its documentation, but the provided facts include no evidence of the stability or performance of device control. The changes in v3.3.0 cannot be confirmed here either, so there is not enough basis to interpret this release as an improvement or an addition of a specific feature. The currently verifiable scope is therefore the project's direction — natural-language mobile app control and information extraction — while the reliability of real-world behavior and the difference of the latest version cannot be judged from this information alone.