visionvision.ocr

OCR

Read text from a screen region with OCR.

Inputs

4
InputValue typeRequirementAccepted sources
languagestringRequiredliteral
pageSegModestringRequiredliteral
resultLevelstringRequiredliteral
ROIroiOCR region. Value type: screen-ratio rect {x, y, width, height}; 0..1 maps to the device frame. Smaller ROI reduces unrelated text.rectRequiredliteral

Outputs

1
Output pathValue typeCoordinate space
data.resultsarray

Runtime support

DesktopSupported
Android AgentSupported