Practical overview
When it fits
Use Mobilerun when the agent needs to interpret mobile UI state and combine actions such as app launching, tapping, swiping, typing, and screenshot analysis. Choose the local framework when the agent should run on your machine; the project points to Mobilerun Cloud for managed infrastructure and API-driven device workflows.
Who it is for
- Developers building LLM-driven mobile automation
- Teams creating mobile UI testing or regression workflows
- Python users who need custom tools around device actions
- Builders automating routine Android or iOS tasks with natural-language instructions
Documented uses
- Issue natural-language commands for mobile-device actions
- Inspect UI trees, screenshots, and device state
- Plan and execute multi-step workflows
- Create custom Python automation with tools and structured output
- Trace agent execution with Arize Phoenix or Langfuse
Requirements & limits
The repository is public, open source, and MIT-licensed. Local installation is documented through uv. The primary setup requires Python >=3.11,
- Android uses a Portal app with ADB, while iOS follows a separate iOS Portal flow.
- The Python API suits applications that need custom tools and workflows beyond one-off CLI commands.
- Vision and vision-only modes support screenshot-based interaction when accessibility-tree information is insufficient.
- Provider choice affects configuration: Anthropic requires the documented optional extra, while other listed providers are included by default.
- The framework is designed for local agent execution; managed device infrastructure is presented separately through Mobilerun Cloud.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 2 October 2026
- Original website title
- droidrun/mobilerun · Official homepage title · retained in its source language

