Generate captions, optimize content, and enhance with custom images for maximum engagement.
Droidrun
Open-source AI platform that lets LLMs directly control Android and iOS devices to automate mobile app interactions via structured UI text and self-correcting loops.

About Droidrun
Droidrun is an open-source AI-driven mobile automation platform that enables large language models to natively control real Android devices and experimentally support iOS. It converts app user interfaces into structured text, allowing LLMs to perform actions, observe outcomes, and re-evaluate steps in a robust Action→Observe→Re-evaluate loop for self-correcting automation. The tool supports multiple LLMs including OpenAI, Anthropic, Gemini, and Ollama, and provides a Python SDK and CLI for integration. Additional features include credential management, parallel device control, macro recording, and workflow orchestration for complex automation sequences. Droidrun is designed for developers and teams who need to automate repetitive mobile app tasks without manual intervention, such as testing, data entry, or UI interactions across multiple devices simultaneously. Its open-source nature allows for customization and extension to fit specific automation needs.
Key features
- LLM-driven control of real Android and experimental iOS devices
- UI-to-structured-text conversion for actionable automation
- Action→Observe→Re-evaluate loop for self-correcting automation
- Support for multiple LLMs (OpenAI, Anthropic, Gemini, Ollama)
- Python SDK and CLI for integration and scripting
- Credential management for secure automation
- Parallel device control for multi-device workflows
- Macro recording and playback for repetitive tasks
- Workflow orchestration for complex automation sequences
Use cases
- Automating mobile app testing across multiple devices
- Data entry and form filling in mobile applications
- UI interaction automation for repetitive workflows
Pros
- Provides real Android devices for AI agents to interact with, enabling authentic app testing and automation
- Supports both cloud-based and local device fleets for scalable automation workflows
- Offers an open framework for customization and extension of mobile automation capabilities
- Enables AI agents to operate with real identities and credentials across applications
- Achieves high performance benchmarks, such as 91.4% on AndroidWorld
Cons
- Experimental support for iOS may not be as mature or reliable as Android functionality
- Requires technical expertise to fully leverage the open framework and customization options
Frequently asked questions about Droidrun
What is Droidrun and what does it do?
Droidrun provides a 'Phone Layer' for AI agents, giving them access to real Android devices to perform tasks such as app testing, data entry, and UI interactions through an API. It enables agents to operate with real identities and credentials in authentic environments.
Who is Droidrun designed for?
Droidrun is designed for developers, teams, and AI agents that need to automate mobile app interactions, particularly for testing, workflow automation, or data processing tasks across multiple devices.
Does Droidrun support iOS devices?
Droidrun experimentally supports iOS devices, but its primary and most reliable functionality is focused on Android. The iOS support may have limitations or require additional configuration.
How can I get started with Droidrun?
Users can start by exploring the open framework or cloud-based device fleets available through Droidrun's platform. Documentation and guides are provided to help integrate AI agents with real devices for automation tasks.
What kind of integrations does Droidrun support?
Droidrun supports integration with AI agents and frameworks via API, allowing them to control real devices for automation. It also provides a Python SDK and CLI for developers to build custom workflows.
Can Droidrun be used for testing mobile applications?
Yes, Droidrun is designed to enable realistic mobile app testing by providing AI agents with access to real devices, allowing them to interact with apps as a human would.
Droidrun traffic & engagement
droidrun.ai · third-party estimates from Similarweb
Traffic trend
Last 4 months · grows by one month every month
- Jun 202611K
- Jul 202621K+87.9%
- Aug 20269.3K−55.6%
- Sep 20268.4K−8.9%
Where visitors come from
Share of visits by channel
Top countries
Share of visits · estimated visits
Compared with similar tools
Automation & Agents · Sep 2026
Traffic figures are third-party estimates, good for scale and trends; they can differ from a company's own numbers. Updated monthly.