Mistral OCR

Mistral OCR is a document understanding model from Mistral AI that goes beyond traditional optical character recognition to provide structured extraction of text, tables, equations, and layout information from documents and images. It handles complex document formats including PDFs, scanned documents, and photographs of text.

The model understands document structure, preserving the hierarchy and relationships between elements like headings, paragraphs, tables, and figures. It supports multiple languages and scripts, handling technical documents with mathematical notation, code blocks, and mixed content types.

Mistral OCR is designed for enterprise document processing, automated data extraction, and digitization workflows where understanding document content and structure is critical.