Mathpix converts PDFs, images, and handwriting to LaTeX, Markdown, DOCX, and Excel. Free tier, Pro from $4.99/month, plus API and enterprise options.
Researched and fact-checked by Bartu Cavusoglu · Updated October 26, 2026
Mathpix is a document conversion service that turns PDFs and images into searchable, exportable, machine readable text. Its OCR models are built for STEM material, including math, chemistry, handwriting, tables, and text in multiple languages. The company says it was founded in 2016 and serves more than 5 million users and over 3,000 organizations. Its stated mission is to make all the world's documents accessible and searchable.
Mathpix sells to two groups. Individuals and teams use Snip, an app for converting images and PDFs, along with a Snipping Tool desktop app and a Chrome extension. Developers and enterprises use the OCR API, Files API, Private Cloud OCR, and the Secure Conversion Service. The homepage reports 50 billion PDF pages and 5 billion images converted to date.
In Snip, a user captures an equation, table, or page from an image or PDF and receives the converted result. The Snipping Tool works from the screen: a keyboard shortcut copies recognized math or chemistry to the clipboard, ready to paste into LaTeX, Markdown, MS Word, and other tools. The conversion gallery shows results in formats such as MMD, LaTeX, TSV, SMILES, and JSON.
Developers reach the same recognition through the OCR API, with documentation, an API reference, and a Python SDK. For large libraries, the Secure Conversion Service uses dedicated capacity that the company says handles millions of pages of unstructured PDFs and images per hour. Mathpix describes this as a way to prepare accurate data for training and tuning language models.
The conversion gallery shows how output keeps document structure. Handwritten notes return as headings and bulleted lists, and multi-line aligned equations stay in a single aligned environment. Subscripts survive as math instead of flattened characters, and source code is kept as a code block with indentation intact. Mixed-script pages, such as Chinese text with inline math, are read in one pass.
Tables come back with merged multi-row and multi-column headers, and two-row headers carrying units stay as two rows. Chemical structures are returned as SMILES strings, and the site lists image to ChemDraw conversion. Snip also includes a PDF reader, PDF search, and PDF data extraction. Exports include DOCX, LaTeX, Overleaf, HTML, Markdown, and Excel.
Mathpix is a document conversion service that turns PDFs and images into searchable, exportable, machine readable text. It offers the Snip app for researchers and students, and OCR APIs and conversion services for developers and enterprises.
Snip has a free plan, Pro at $4.99 per month ($49.90 billed yearly), and Organization at $9.99 per month ($99.90 yearly) with two users included. Extra usage costs $0.0035 per PDF page and $0.002 per image. Enterprise pricing is custom.
Yes. The Free and Educational plan includes 10 PDF pages and 10 images per month. Users who sign up with a .edu email get twice that usage.
The site lists conversion to LaTeX, Markdown, DOCX, MS Word, HTML, Excel, Overleaf, CSV, AsciiMath, ChemDraw, and SMILES. The conversion gallery also shows TSV and JSON output.
Yes. Mathpix lists handwriting recognition for printed or cursive writing. Its gallery shows handwritten lecture notes returned as headings and bullet lists, and handwritten equations returned as inline LaTeX.
Yes. Mathpix provides an OCR API, a Files API, documentation, an API reference, and a Python SDK. Private Cloud OCR and the Secure Conversion Service are available for enterprise needs.
Yes, at any time. The subscription stays active until the end of the current billing cycle and then does not renew. Subscriptions are non-refundable, so cancel before the renewal date.
The site lists downloads for MacOS, Windows, Linux, the App Store, Google Play, and Huawei App Gallery. It also offers a web app and a Chrome extension.