tesseract-ocr/tesseract is the main repository for the Tesseract OCR engine, and the latest publicly available release is 5.5.3. Tesseract is an open source OCR engine that reads characters in images and converts them into text.
The repository provides the OCR engine libtesseract and the command-line program tesseract. Its README states that it supports both an LSTM (Long Short-Term Memory) based OCR engine and the legacy engine of Tesseract 3, and that it can handle several image formats including PNG, JPEG and TIFF. Recognition results can be exported in formats such as text, hOCR, PDF, TSV, ALTO and PAGE, and C and C++ APIs are provided as well. The primary language is C++, and the license is Apache-2.0. The repository was created on August 12, 2014 and has 77087 stars.
However, since the source material contains no release notes, there is no way to know what changed in 5.5.3. There are no measured figures for accuracy or processing speed either, and the only rating that could be confirmed is a Momentum score of 4.3 out of 10. The Apache-2.0 license and the wide range of output formats are favorable conditions for plugging into a document processing pipeline, but until the changes are confirmed through release notes there is not enough basis to adopt it anew. For teams already using it, this is closer to a decision about whether to keep the current version or move up.