Mistral has introduced its Optical Character Recognition (OCR) tool, named Mistral OCR. This tool is designed to enhance document understanding through an advanced API.

The tool processes images and PDFs, extracting media, text, tables, and equations at speeds of up to 2,000 pages per minute on a single node.

Supporting multiple languages and structured JSON outputs, Mistral OCR enables seamless integration into technical workflows and applications.

For organizations with strict data compliance regulations, Mistral offers a self-hosting option to keep sensitive documents within internal systems.

Mistral OCR is available on Mistral’s developer platform, with a pricing structure of US$1 for every 1,000 pages.

Recent Mistral developments