Frames group cards visually. The AI used to be blind to that grouping, so asking it to add something to a frame was a matter of luck.

What changed

The AI now reads frame membership and writes it:

• It knows which cards a frame contains
• A card it makes inside a frame is captured by that frame
• A captured card is positioned relative to its frame, and moves when the frame moves

Edge connections and frame dragging follow the same rule, so nothing drifts out of a frame when you reposition it.

Asking for something inside a frame

Name the frame in your request. "Add three more shots inside the act-two frame" lands inside it, rather than nearby and unattached.

Use Cases

• Ask for three more shots inside the act-two frame and have them land inside it
• Move a frame across the board and take everything the AI put in it along