← 中文版本

MacroAI

A next-generation cross-platform macro automation platform (Windows / Android).

Core positioning: Visual node-based workflow + Intelligent execution + Cross-platform reusable scripts. Developers can create automation scripts with MacroAI and distribute them to other users. The same Lua scripts and image templates run on both Windows and Android. Built-in multi-resolution adaptation ensures the same script runs stably across different screen sizes and scaling settings.

→ Website · User Guide · API Reference

---

📥 Download

Windows

Microsoft Store: Microsoft Store GitHub: Releases Gitee: Releases

Android

GitHub: Releases (direct APK download) Gitee: Releases (direct APK download)

---

✨ Key Features

Cross-Platform Scripts · Write Once, Run Anywhere

The same Lua script runs on both Windows and Android, with image templates shared across platforms. Scripts generated by the Windows visual node editor run directly on Android.

Visual Node Editor · No-Code / Script Dual Mode (Windows)

Drag-and-drop tree-container architecture with 20+ node types. LOOP and CONDITION act as container nodes that nest child nodes, showing the flow hierarchy visually. All node operations automatically generate Lua code — approachable for non-programmers, extensible for developers.

Background Input Simulation · Zero Focus Stealing (Windows)

Leverages Win32 PostMessage/SendMessage APIs for background mouse and keyboard operations. Scripts execute silently in the background while you continue using your computer uninterrupted.

OCR Text Recognition · No Model Download Required

The Windows edition uses the built-in Windows.Media.Ocr (WinRT) API, while the Android edition uses ML Kit OCR. Neither requires model file downloads — Chinese OCR works out of the box.

Screen Change Detection · Smart Filtering

Grid-based comparison with temporal filtering distinguishes meaningful changes from periodic flickering (cursors, animations). Bar segment color tracking monitors health bars, progress bars, and similar visual indicators.

Multi-Resolution Adaptation · Cross-Device Scripts

Uniform/desktop dual-mode coordinate transformation with DPI awareness. The recording design area is automatically mapped to the runtime actual area. The same script automatically adapts to different resolutions and scaling ratios — the core enabler for script reusability and distribution.

Accessibility Service Driven · No Root Required (Android)

Built on Android AccessibilityService for automated gestures such as taps and swipes — no Root access needed. Screen capture permission is requested on-demand only when image recognition is required.

Multi-Threaded Architecture · Smooth Performance

Scripts execute in a dedicated thread without blocking the UI. ActionManager supports concurrent execution of multiple custom actions. Audio and dialogs are bridged back to the main thread via Qt Signal.

IME Input Method Compatibility · Accurate Chinese Recording (Windows)

Polls the focused control via WM_GETTEXT diff during recording to capture IME text, rather than recording raw key events. Ensures accurate Chinese text capture during input method composition.

Bilingual Interface · Full i18n

Complete Chinese and English UI, with matching help documentation and API reference in both languages. Switch between them in settings with one click.

Completely Free · No Ads

All core features are permanently free — no feature gating, no ad popups, no bundled adware.

---

System Requirements

Windows

• Windows 10 / Windows 11 (64-bit)
• 4GB+ RAM
• 200MB+ disk space
• DirectX 11 compatible graphics card

Android

• Android 8.0 or later
• Accessibility Service permission required (for automation)
• Screen capture permission required for image recognition
• ~30MB storage space

Tech Stack

Windows Edition

• GUI: PySide6
• OCR: Windows OCR (WinRT)
• Image Processing: OpenCV
• Scripting: Lua (lupa)
• System Interaction: pyautogui (foreground), Win32 API (background)
• Recording: pynput
• Audio: QMediaPlayer, winsound

Android Edition

• GUI: Jetpack Compose
• OCR: ML Kit OCR
• Image Processing: OpenCV Android SDK
• Scripting: Lua (luaj)
• Automation: AccessibilityService + MediaProjection