Mathpix OCR API

The Mathpix OCR API converts documents into structured, editable output — preserving complex layouts, inline images, and content that general-purpose tools miss.

Features

Input formats

• Documents (e.g. .pdf, .docx), e-books, and more
• Images (JPEG, PNG, BMP, WebP, TIFF, and more)
• Handwriting stroke coordinates
• Batch processing for multiple images

Recognized content

• Printed and handwritten text
• Tables with cell spanning
• Math equations (LaTeX, MathML, AsciiMath)
• Per-line confidence scores
• Chemical structures (SMILES)

Output

• Mathpix Markdown, Markdown, office documents (.docx, .pptx), LaTeX, HTML, and more
• Page-level streaming (SSE) for long documents
• Auto-rotation correction

Result management

• Query and retrieve past results, filter by date or tags
• Delete PDF and conversion results when no longer needed
• Monitor API consumption
• Privacy controls for data retention

Language support: English, Chinese, Japanese, Korean, Hindi, Russian, Thai, Vietnamese, Tamil, Telugu, Gujarati, Bengali

Getting started

Follow the guides to make your first request.

Support

Questions or problems? Email support@mathpix.com.