Distribution & Import/Export

MacroAI can export a project as a single file that others can import and use directly:

1. Open your project in MacroAI
2. Click the menu Project → Export Project
3. Choose a save location to generate a .macroai file
4. Send it to the other person via email, messaging apps, or any other method

After they receive it:

• Windows: double-clicking the .macroai file automatically opens MacroAI and imports it (if installed)
• Or in MacroAI, click Project → Import and select the file
• Android: open the file with MacroAI — it's imported automatically and takes you to the run screen

The exported file contains the complete project data: all nodes, images, variables, functions, and more.

When exporting, you have two options:

• Normal export — a plaintext package. After import, all nodes are visible and editable. Suited for: backups, team collaboration, switching computers yourself.
• Distribution package (encrypted export) — the script content is encrypted. After import, it can only be run; nodes can't be viewed or edited. Suited for: distributing to end users, protecting your design work.

Characteristics of a distribution package:

• After import, it automatically enters run mode; the design mode entry is grayed out and can't be clicked
• Users can't see the node editor, code, or other design content
• Users can only: calibrate the region, fill in personal configuration, and run/stop
• Encryption-protected — the script can't be recovered by unpacking the file

Check the "Distribution package" option when exporting to generate an encrypted package.

That's because what you imported is a distribution package (encrypted package). The whole purpose of a distribution package is "let users run it, but not view/modify the script".

After importing a distribution package:

• It automatically switches to run mode
• The design mode entry is grayed out and can't be switched to
• You can only see: project info, region calibration, personal configuration, and the run button

This is expected behavior, not a bug. If you need to edit the script, ask the designer for a normal export package (unencrypted), or have them give you the source project.

After importing, switch to run mode. Usually you'll need to:

1. Calibrate the run region (if your screen resolution/window position differs from the designer's)
   • Click the "Calibrate" button and select the actual position of the target program on your screen
   • If the designer used full screen and you're also running full screen at the same resolution → you can skip this
2. Fill in the personal configuration (if the designer defined user config variables)
   • E.g. upload your own avatar image, or enter your in-game nickname
   • The interface shows the items that need to be filled in
3. Click Run

The run mode screen shows a "configuration completeness" indicator telling you which settings are still missing.

User config variables let the designer reserve "customizable slots" that users fill in with their own values before running.

Typical scenario: the script needs to click the user's in-game avatar, but everyone's avatar is different → the designer defines an "avatar" image variable, and the user captures their own avatar before running.

Designer's steps:

1. Define the variable in the "User Config" tab (Ctrl+9) (e.g. ID=avatar, Name=My Avatar, Type=Image)
2. In the Find Image node, reference that variable via the ◎ button (the field shows u|avatar)

User's steps:

1. See the "My Avatar" config item on the run mode screen
2. Click capture/upload to provide their own avatar image
3. At runtime, the script automatically uses the image the user provided

Yes. Both platforms use the same project format and encryption protocol, and .macroai files are compatible in both directions.

A script built on Windows can be exported and imported to run on Android, and vice versa.

Things to note:

• Coordinate strategy and resolution adaptation still need to be handled (the two platforms have different screen sizes, so calibrate before running)
• Some nodes may be supported on one platform but not the other (unsupported nodes are skipped after import)
• Windows "Key Press" nodes use virtual key codes, which may behave differently on Android