Framework · open source

AI mobile automation framework for iOS and Android.

Open-source framework for AI agents that control iOS and Android devices. LLM-agnostic. Multi-agent architecture.

Star on GitHub → MIT Licensed · Python 3.10+ · Android 10+

agent.py

1import asyncio
2from mobilerun import MobileAgent, MobileConfig
3 
4async def main():
5    config = MobileConfig()
6    agent = MobileAgent(
7        goal="Open Twitter and post 'Hello from mobilerun!'",
8        config=config,
9    )
10    result = await agent.run()
11 
12asyncio.run(main())

LLM-agnostic

Works with any LLM.

Swap providers in one line. OpenRouter wired in for the rest.

Gemini

OpenAI

Anthropic

DeepSeek

Qwen

Ollama

OpenRouter

Showcase

Watch it in action.

See mobilerun control real Android devices with natural language.

Agent archetypes

Automate anything on mobile.

Reliably and effortlessly. Three archetypes for every workflow.

01 · Task Agents

Task Agents.

Automate complex multi-app workflows through natural language.

02 · Testing Agents

[ ]

Testing Agents.

Write mobile tests in plain English — no Appium, no XPath.

03 · Research Agents

Research Agents.

Extract structured data from any mobile app.

Community

Built in the open.

Benchmark

91.4% on AndroidWorld.

State-of-the-art performance on the leading public benchmark for mobile agents. 106 of 116 tasks completed. View benchmark →

Success rate 91.4%

106 of 116 AndroidWorld tasks completed.

Primitives

Everything you need
to ship.

Six primitives covering the full mobile automation lifecycle.

LLM Agnostic

Swap providers with one line. Gemini, OpenAI, Anthropic, Ollama, DeepSeek, OpenRouter, Qwen and more.

Vision Mode

Screenshot-based fallback when the accessibility tree is insufficient. Combine both for maximum accuracy.

Reasoning Mode

Multi-step planning with chain-of-thought. The agent reasons about UI state before acting.

Structured Output

Extract typed data with Pydantic models. Get real objects, not raw strings.

Custom Tools

App Cards

Pre-built interaction patterns for popular apps. Skip the boilerplate, start automating.

Local → cloud

Start local. Scale to cloud.

Same agent framework. Run it on your own device, or on the managed cloud.

Open source

mobilerun framework

pip install
Local devices
Full control

Shared protocol SDK · trajectories · prompts

Cloud scale

mobilerun cloud

Hosted devices
Parallel agents
Stealth mode

Capability	Legacy automation	Mobilerun framework	Mobilerun cloud
LLM-driven workflows	No	Yes	Yes
Device source	Manual	BYO device	Managed fleet
Mobile-integrity readiness	No	DIY	Included
Fleet scale	No	Self-host	1 → 10,000
iOS	No	Limited	Full
eSIM & SMS	No	No	Yes

Scale to the cloud →

Quickstart

Zero to agent in 60 seconds.

Three commands. One device. Full automation.

terminal

$ pip install mobilerun
Successfully installed mobilerun-0.5.8

$ mobilerun setup
Portal installed and accessible. You're good to go.

$ mobilerun run "Open Settings and check battery level"
[agent] planning steps…
[step 1/2] tap "Settings"
[step 2/2] navigate to "Battery"
[done] task completed in 3.8s — 2 actions, 0 retries

Open source.
Production runtime.

Build on the framework. Ship on the cloud. Same agent, real-device identity.

Star on GitHub → Cloud pricing