Sorry, as added questions: Does the API return structured data per identified food item, or a single aggregate for the whole plate? What is the confidence/accuracy model — does it return uncertainty ranges? Is there a session number or a framework name I can search in the Developer documentation?

Thank you for the encouraging answer. Is there any public documentation about this API to receive the visually extracted meal nutritional content? Is there a WWDC session on the vision framework related to extracted male nutritional content?

Addendum: Does Xcode 27 developer beta with Swift 6.4 fix the MainActor bug in the Swift compiler?