OptionalDescriptionA 1-3 sentence summary of what is visible/actionable, or null when nothing could be interpreted.
OptionalDetailNon-fatal note explaining a null/empty result (e.g. 'no live browser', a parse/vision error).
UI elements matching the request, each with a pixel centroid. Empty for a describe-only request.
Result of RemoteBrowserActionResolver.InterpretRemoteBrowserPage — how a fast vision model "sees" the current browser viewport on behalf of a voice agent that cannot view images.
Descriptionis a concise text summary;Elementslocalizes any UI elements the request asked for (empty for a plain describe).Detailcarries a non-fatal note (e.g.'no live browser'or a vision-error message) — every path is best-effort and never throws.