Feature Request: Customizable PiP Size (width/height) in OpenAI Apps SDK
Summary
I would like to request a feature that allows developers to customize the width and height of the Picture-in-Picture (PiP) mode in the OpenAI Apps SDK. Currently, PiP is rendered with a fixed layout controlled by the host, and there is no API to adjust its dimensions. This limitation significantly reduces UI/UX flexibility for mini apps.
Current Behavior
When calling:
await openai.requestDisplayMode({ mode: "pip" });
The app switches to PiP mode with a predefined size. The PiP window often occupies a large portion of the screen (sometimes close to half of the viewport). Because its size is fixed, the PiP frequently blocks underlying ChatGPT UI elements, making interactions difficult. Developers cannot control or adjust its width or height programmatically, which makes PiP hard to use in real workflows.
There is no documented or undocumented API that allows resizing or configuring the PiP container.
Expected / Desired Behavior
Provide an API that allows developers to optionally specify PiP dimensions, for example:
await openai.requestDisplayMode({
mode: "pip",
width: 360,
height: 240
});
Or an equivalent preset-based or constraint-based configuration that enables more compact or adaptive PiP layouts.
Thank you for considering this feature request.
