Select on the original.
The inspector lets you click a detected face or person on the original thumbnail. Choose a subject directly instead of guessing a crop position.
Face-aware crops, person-aware framing and a clear way to choose the subject that matters to your layout.
The inspector lets you click a detected face or person on the original thumbnail. Choose a subject directly instead of guessing a crop position.
See the actual crop outline. Group-preserving framing aims to keep selected subjects together when the requested aspect ratio and source geometry allow it.
Use a tighter face portrait, add context with facepad, or place a manual focal point and control zoom. Your choices become URL parameters.
An explicit explanation reports when no matching subject is found or a group cannot fit. Adjust the crop rather than assuming every subject was preserved.
Detection analysis is cached for reuse across output sizes, with bounded analysis quotas. Creating another output size does not inherently require another analysis.
Detection returns face and person locations. It does not recognize who someone is or build an identity database.
The image API covers automatic focus, explicit subject selection and manual framing.
?w=640&h=360&fit=crop&crop=faces?w=320&h=320&fit=facearea&faceindex=1&facepad=1.5Detection can miss a subject. An aspect ratio can exclude part of a group. Review representative images and the crop explanation before using the result. Output is not promised to be pixel-identical to another provider.
Let’s build a simpler image pipeline for your next million requests.
Enterprise image optimization. Built for scale.