SuperAPI AI Directory

replicant-mcp

Android development: build APKs, launch emulators, manage devices, automate UI and analyze logs.

Open website

About this tool

replicant-mcp is a Model Context Protocol (MCP) server for Android development that enables AI assistants (such as Claude) to interact with the Android development environment through natural language. Its main purpose is to help AI build, test, and debug Android apps, covering the complete workflow from building APKs, launching emulators, installing apps, and browsing the UI to debugging crashes.

Main capabilities include: custom build commands (supporting project-specific overrides and automatic detection of gradlew), video capture (start/stop recording, duration-based capture), physical device and emulator management, device operations via adb, and accessibility-first UI automation. The tool uses a progressive disclosure mechanism, providing summaries first and details on demand to minimize token usage; at the same time, accessibility-first UI automation is faster, more economical, and more reliable than screenshot-based approaches.

Usage: Requires the Android SDK (including adb and emulator) and ensures it is in PATH, as well as an Android project containing gradlew. If the tool is not installed, it can be installed via Homebrew. When using only a physical device, only adb is needed; adb will be added directly to PATH with no additional configuration required. The full Android SDK (for emulator workflows or building APKs via the gradle-* tools) requires executing the installation steps in order and persisting ANDROID_HOME and appending it to PATH in ~/.zshrc or ~/.bashrc. After installation, run the built-in diagnostics to verify the environment. Configuration: For Claude Desktop, edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows); for Cursor, edit ~/.cursor/mcp.json (global) or .cursor/mcp.json (project); for Windsurf, edit ~/.codeium/windsurf/mcp_config.json; for OpenAI Codex CLI, edit ~/.codex/config.toml. Ensure that adb and emulator are in PATH, or set the ANDROID_HOME environment variable (macOS: ~/Library/Android/sdk; Linux: ~/Android/Sdk).

Configuration and logging: Advanced options can be set via REPLICANT_CONFIG; server logs can be controlled via REPLICANT_LOG_LEVEL (error, warn, info, debug) and REPLICANT_LOG_FORMAT (json for structured output), with logs written to stderr. For troubleshooting, see docs/troubleshooting.md; tool documentation allows Claude to call rtfm and specify categories such as "build", "adb", "emulator", or "ui". For development setup and guidelines, see CONTRIBUTING.md.

Use cases: Developers who need AI assistance for Android app building, testing, UI automation, and debugging, especially scenarios where they want to reduce token consumption and improve the reliability of UI interactions.

Limitations: The tool requires an Android SDK environment (adb, emulator) and an Android project containing gradlew; some features (such as emulators and building APKs) require the full SDK. Information about fees, pricing, platform support, etc. is not explained in the provided materials and is not currently available.