OCR
mash.ai.ocr · AI
OCR processing and miniOCR format conversion
Methods
Section titled “Methods”compact
Section titled “compact”ocr.compact(miniOcr: string) → stringRender miniOCR as full compact format: x,y,w,h:word Preserves all bounding box data for coordinate mapping.
positional
Section titled “positional”ocr.positional(miniOcr: string) → stringRender miniOCR as compact positional format: |y| x:word x:word Preserves x/y coordinates without per-word dimensions.
spatial
Section titled “spatial”ocr.spatial(miniOcr: string, columns?: int = 120) → stringRender miniOCR as fixed-width spatial text preserving visual layout. Best for general LLM extraction (invoices, receipts, forms).
ocr.text(miniOcr: string) → stringStrip all XML tags and coordinates, returning plain text.