Open Internet by MindsNet

datalab-to/chandra

Chandra OCR is a state-of-the-art model that converts images and PDFs into structured HTML, Markdown, or JSON while preserving layout information. It's designed for document intelligence tasks, handling complex tables, forms, handwriting, and multilingual text. With its advanced features and high performance, Chandra OCR is a top choice for developers and organizations seeking accurate document processing. Best for: Developers and organizations seeking advanced document intelligence and OCR capabilities Use cases: Automating document processing and data extraction from invoices, receipts, and forms; Converting scanned documents and images into structured digital formats; Building AI-powered document analysis and understanding applications Highlights: Handles complex tables, forms, handwriting, and multilingual text with high accuracy; Supports over 90 languages and has excellent performance on internal multilingual benchmarks; Offers two inference modes: local (HuggingFace) and remote (vLLM server) for flexible deployment

Computing & Technology, Computer Science, Software Engineering