Type Alias: VisionContentCategory

VisionContentCategory = "printed-text" | "handwritten" | "document-layout" | "photograph" | "diagram" | "screenshot" | "mixed"

Defined in: packages/agentos/src/core/vision/types.ts:57

What kind of visual content the pipeline detected or was told to expect.

Used to route images to the most appropriate processing tier — e.g. handwritten content triggers TrOCR, complex layouts trigger Florence-2.