OCR
Extract and structure text from images and documents
OCR models extract text from images, scanned documents, PDFs, and photos. Modern OCR models preserve document structure, handle tables, detect handwriting, and output structured formats like Markdown or HTML.
Available Models
- Nanonets OCR2 3B – Advanced OCR with structured Markdown output, signature/watermark detection, and handwriting support
- PaddleOCR-VL – Vision-language OCR with handwriting, table-to-HTML conversion, and prompt-based extraction