Mobile-use automates control of Android and iOS devices via natural language. The project provides a Python-based agent framework with platform-agnostic configuration and Docker/dockerless usage paths.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
Mobile-use is an open-source AI agent framework that controls Android or iOS devices using natural language. It supports UI navigation, data scraping, and extensible LLM configurations for various providers.
How it works
The README describes using a modular agent system that can interact with mobile apps by interpreting natural language and performing UI actions. It exposes commands through Python scripts and a Docker quickstart for Android devices, with instructions to configure environment variables and LLM providers. It supports multiple models/providers via a configurable llm-config.override.jsonc and an llm-config.defaults.jsonc preset.
Getting started
From source: "cp .env.example .env" and set keys. "cp llm-config.override.template.jsonc llm-config.override.jsonc" to customize LLM config, then edit accordingly. To run via quickstart (Docker):
The README provides a detailed Quick Launch (Docker) section with platform-specific command blocks for Linux/macOS and Windows, including mobile-use.sh/mobile-use.ps1 invocation.
Installation steps:
-
Clone the repo: "git clone https://github.com/minitap-ai/mobile-use.git && cd mobile-use"
-
Create and activate a virtual environment using uv: "uv venv"
-
Install dependencies: "uv sync"
Usage examples: "python ./minitap/mobile_use/main.py "Go to settings and tell me my current battery level""
Recent releases
Latest releases include:
- v3.3.0 (2026-01-12): fixes input normalization for press key tool and adds form filling guidelines.
- v2.9.0 (2025-11-15): adds support for naming tasks for tracing and clarifies API key/config in pla.
- v2.6.0 (2025-10-20): mentions bumping to v2.5.0 and iOS device detection fix using Darwin.
- v2.5.3 (2025-10-11): notes similar changes as v2.5.2/v2.5.0.
- v2.5.2 (2025-10-11): same notes as v2.5.3.
Traction
Stars: 2727; Forks: 232; Open issues: 3. Language: Python. License: Apache-2.0. Created: 2025-08-16. Last push: 2026-07-09.
Behind the repo
No linked startup/company information is provided in the facts.
Caveats
License: Apache-2.0. Age: created 2025-08-16. The README notes platform support and device types, with certain notes about iOS physical devices not yet supported and hardware dependencies (ADB, Xcode, fb-idb).