Describe the Feature Request
When interacting with multimodal models, the underlying model supports image inputs. However, within the Antigravity CLI's terminal interface (TUI), there is currently no way to directly paste an image from the system clipboard using Ctrl+V (or Cmd+V).
Even though the underlying backend code is designed to accept image inputs, the terminal frontend cannot capture the binary clipboard paste action.
Proposed Behavior
- When a user presses Ctrl+V (or Cmd+V) and the clipboard contains an image:
- The CLI captures the clipboard image data.
- It automatically saves the image to a temporary file.
- It inserts the file path of the saved image into the prompt bar.
- If direct clipboard reading is restricted by the terminal environment, provide a clear tooltip or error message suggesting the user drag-and-drop the file instead.
Current Workarounds
- Manually saving the screenshot to a file, locating the path, and typing/dragging the path into the CLI input.
Additional Context
- The backend engine and model integrations have the necessary code/support to receive visual context, but the terminal user interface (TUI) acts as a bottleneck because it does not bind clipboard image access to paste actions.
Describe the Feature Request
When interacting with multimodal models, the underlying model supports image inputs. However, within the Antigravity CLI's terminal interface (TUI), there is currently no way to directly paste an image from the system clipboard using Ctrl+V (or Cmd+V).
Even though the underlying backend code is designed to accept image inputs, the terminal frontend cannot capture the binary clipboard paste action.
Proposed Behavior
Current Workarounds
Additional Context