Repository Issue Activity (beta)

#4619 inputs from imagelist or stdin with filenames starting with 'MM' misdetected as tiff

#4618 lstmtraining --convert_to_int leaves the ADAM training_flags_ bit set, unlike official int-optimized traineddata

#4615 Incorrect ALTO word confidence conversion

#4610 Please enable Private Vulnerability Reporting (security disclosure channel)

#4595 Regression: Tesseract 5.5.3 recognizes uppercase 'I' as '|' while 5.5.0 recognizes it correctly

#4593 Downstream integration: contributors wanted for Tesseract OCR support in Heym

#1714 [Feature Request] Table structure extraction at the API

#4591 write runtime parameters to HOCR output

#4589 Training unicharset_extractor broken in 5.5.3

#4504 Installation binaries are missing from releases page

#4498 Plans for Tesseract in 2026

#4587 [ CRITICAL BUG ] Request: private security-vulnerability reporting

#4525 Better documentation for --print-parameters.

#4582 textord noise filter silently drops entire text lines in dotted Arabic-script languages (verified: Persian, Arabic, Urdu); textord_noise_rejrows=0 recovers them

#4580 Request: enable Private Vulnerability Reporting / security contact

#3446 Page numbers not detected in various cases

#4523 OCR accuracy regression after RNG revert in 5.5.1: LSTM hallucinates extra characters

#2738 Duplicate Characters in Output Stream

#4448 Crash upon processing input file with disabled legacy engine and custom traineddata

#4544 Arabic words are sometimes reversed in OCR output

#4458 LSTM models recognize random characters instead of asterisk (*)

#4430 Tesseract takes much time when processing image

#4443 lstm can not be generate by --psm 6 lstm.train

#4549 Evaluate std::simd / std::experimental::simd

#4552 libtesseract.la needs '-framework Accelerate' to harden build