Command Palette

Search for a command to run...

UnylyUnyly
Browse all

Mobile Agent Platform

FreeNot checked

MCP server for operating and testing Android devices with AI agents. Provides observation and action tools for interactive control and spreadsheet-driven test r

GitHubEmbed

About

MCP server for operating and testing Android devices with AI agents. Provides observation and action tools for interactive control and spreadsheet-driven test runs.

README

An open, extensible platform for operating and testing Android devices with AI agents.

中文说明 · Contributing · Roadmap · Security

Status: v0.1.0 source preview. APIs and plugin interfaces may change before v1.0.

What it provides

  • A local web workspace for interactive device operation.
  • An MCP server exposing Android observation and action tools.
  • Spreadsheet-driven, repeatable mobile test runs with evidence and reports.
  • Claude-compatible model provider settings for Anthropic, Kimi-compatible, and custom endpoints.
  • A small plugin API plus a built-in example plugin.
  • Source scripts for building the optional Windows companion. No prebuilt binaries are committed.

Requirements

  • Python 3.10–3.12 (Python 3.12 x64 for the Windows companion build)
  • Android SDK Platform-Tools (adb)
  • An Android device with USB debugging enabled
  • Credentials for a supported Claude-compatible model provider

Quick start

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
python -m pip install -e ".[dev]"
mobile-agent-platform-web

Open http://127.0.0.1:8000, connect one Android device, and configure the model provider in Settings. API keys are kept in process memory and are not written by the application.

For MCP-only use:

mobile-agent-platform-mcp

Model providers

The current agent runtime uses a Claude-compatible adapter. The UI offers:

  • Anthropic
  • Kimi-compatible preset
  • Custom Claude-compatible endpoint

Provider selection is an adapter boundary, not a platform identity. Contributions that add portable providers without coupling core device tools to a vendor are welcome.

Plugins

Plugins can contribute agent tools and optional settings routes through the mobile_agent_platform.plugins entry-point group. See mobile_agent_platform/plugins/example for the minimal built-in example.

Spreadsheet test suites

The runner reads .xlsx suites and produces evidence-rich run output. A suite normally contains a sequence number, operation, observation point, and expected observation. Generate a neutral example workbook with:

python examples/create_example_suite.py

Windows companion

The repository contains source build scripts under mobile_agent_platform/companion. Download and verify the official Windows Android SDK Platform-Tools separately, then run the PowerShell build script on Windows 10/11 x64. See the Windows build guide.

Community

Please read CONTRIBUTING.md before opening a pull request. Community conduct follows CODE_OF_CONDUCT.md. Governance and maintainer responsibilities are described in GOVERNANCE.md.

License

Licensed under the Apache License 2.0. See NOTICE and THIRD_PARTY_NOTICES.md.

from github.com/mobile-agent-platform/mobile-agent-platform

Installing Mobile Agent Platform

This server has no published package — it is built from source. Open the repository and follow its README.

▸ github.com/mobile-agent-platform/mobile-agent-platform

FAQ

Is Mobile Agent Platform MCP free?

Yes, Mobile Agent Platform MCP is free — one-click install via Unyly at no cost.

Does Mobile Agent Platform need an API key?

No, Mobile Agent Platform runs without API keys or environment variables.

Is Mobile Agent Platform hosted or self-hosted?

Self-hosted: the server runs locally on your machine via the install command above.

How do I install Mobile Agent Platform in Claude Desktop, Claude Code or Cursor?

Open Mobile Agent Platform on unyly.org, pick your client tab (Claude Desktop, Claude Code, Cursor) and press Install — the config is generated automatically, no JSON editing.

Related MCPs

Compare Mobile Agent Platform with

Not sure what to pick?

Find your stack in 60 seconds

Author?

Embed badge for your README

Browse similar

All ai MCPs