tesseract

Tesseract is the OCR engine that's been quietly reading text off images for decades. Point it at a scan or a photo and it hands back the text, in more than 100 languages, as plain text, searchable PDF, hOCR, TSV, ALTO or PAGE. It's Apache 2.0 and free, and it sits underneath a lot of other OCR tools.

It's refreshingly light. Tesseract runs on CPU with no GPU, as a command-line tool or a C++ library that others have wrapped for most languages. Version 5 uses an LSTM neural engine, and it's stable and actively maintained. By default it segments the page into blocks, paragraphs, lines and words before reading, and the structured outputs keep that layout. What it won't do is rebuild tables or pull key-value pairs out of forms.

This is the right tool when you need reliable text extraction and not much more. Solo to large teams: free and dependable. For tables, forms or complex documents, pair it with docling or PaddleOCR, or reach for a heavier model.

The catch is image quality. Tesseract binarizes images itself, but its line detection falls apart on skewed pages, and uneven backgrounds and low contrast hurt accuracy. Budget time to deskew and clean messy scans first.