xby-image-detect

Running object detection inside a conversation workflow typically means standing up a local inference stack first: pulling model weights, configuring a GPU, juggling dependency versions. xby-image-detect collapses all of that into a single API call. No local runtime, no GPU setup, no environment management — hand the image to the plugin and get structured detection results back immediately.

Built on the YOLO26 model, the plugin covers all 80 COCO object classes spanning people, vehicles, animals, and everyday household items. Every detection run returns precise bounding-box coordinates, confidence scores, and class labels that downstream code can consume directly. On the input side it accepts images via URL, Base64-encoded strings, or local file paths, so whether the image lives on a remote server, arrives as an upload payload, or sits on your disk, it plugs in seamlessly.

If your multimodal pipeline needs a lightweight visual-structuring step, you want to quickly confirm which objects appear in a frame, or you simply do not want to be distracted by inference infrastructure setup and maintenance, this plugin lets you skip all the groundwork and focus your energy on the logic that actually drives business value.