Mathpix
Freemium
OCR for math, science and tables that converts images and PDFs to LaTeX and markdown.
About PaddleOCR
PaddleOCR is an open-source toolkit built on the PaddlePaddle framework. It detects and reads text in images in many languages, with models that are small enough to run on ordinary hardware and even mobile devices.
Newer releases add document parsing, which finds the structure of a page including tables and formulas, making it useful for turning documents into data for language models. Pretrained models are easy to download and fine-tune.
PaddleOCR is free under an open license. Developers choose it when they want strong accuracy without cloud fees.
Standout features
- Open source
- Multilingual text recognition
- Document structure parsing
- Table and formula support
- Lightweight models
- Fine-tuning
Tags
More in OCR
View all →Tesseract
Free
The classic open-source OCR engine that recognizes text in over 100 languages.
Mistral OCR
Paid
An OCR model from Mistral that converts PDFs and images to markdown, including tables and equations.
ABBYY FineReader
Paid
A long-standing OCR and PDF editor that turns scans into searchable, editable documents.
Microsoft's service for reading documents and extracting text, tables and fields.
Google Document AI
Paid
Google Cloud's platform for extracting and understanding data from documents.
