#3838th of 69 in Data and scraping for AIFor data and documents, documents and ocr

tesserocr

Python wrapper around the Tesseract OCR C++ API, Pillow-friendly

Stars
2.2k
License
MIT
Last commit
Nov 2026
Last release
Nov 2026
Language
Python
Stars, last 29 weeks+34
25 Apr 2026: 2,148 stars2 May 2026: 2,149 stars9 May 2026: 2,151 stars16 May 2026: 2,153 stars23 May 2026: 2,153 stars30 May 2026: 2,155 stars6 Jun 2026: 2,155 stars13 Jun 2026: 2,156 stars20 Jun 2026: 2,157 stars27 Jun 2026: 2,158 stars4 Jul 2026: 2,158 stars11 Jul 2026: 2,160 stars18 Jul 2026: 2,161 stars25 Jul 2026: 2,161 stars1 Aug 2026: 2,163 stars8 Aug 2026: 2,166 stars15 Aug 2026: 2,167 stars22 Aug 2026: 2,167 stars29 Aug 2026: 2,167 stars5 Sept 2026: 2,170 stars12 Sept 2026: 2,170 stars19 Sept 2026: 2,170 stars26 Sept 2026: 2,170 stars3 Oct 2026: 2,173 stars10 Oct 2026: 2,175 stars17 Oct 2026: 2,177 stars24 Oct 2026: 2,177 stars31 Oct 2026: 2,181 stars7 Nov 2026: 2,181 stars14 Nov 2026: 2,182 stars

2,148 on 25 Apr 20262,182 on 14 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

tesserocr binds directly to Tesseract's C++ API through Cython and exposes it to Python as PyTessBaseAPI plus helpers like image_to_text and file_to_text. It accepts PIL images or file paths, releases the GIL during recognition, and supports free-threaded CPython. Pre-built wheels for Linux, macOS and 64-bit Windows bundle Tesseract and Leptonica.

Who it is for: Python developers adding Tesseract OCR to applications or pipelines

Strengths

  • Calls the C++ API directly via Cython, avoiding subprocess calls to the tesseract CLI
  • Releases the GIL during recognition, so threads run OCR concurrently
  • Wheels bundle Tesseract and Leptonica; no separate install needed
  • Exposes layout iterators, word confidences, orientation/script detection and symbol choices

Weaknesses

  • Language .traineddata files must be installed separately and match the Tesseract version
  • Each worker needs its own PyTessBaseAPI instance; instances have no internal lock
  • Some Tesseract parameters are process-wide and cannot change during parallel OCR
  • Source builds need libtesseract, libleptonica and Cython >=3.1.8,<3.4

What it needs

  • no GPU
  • Docker
  • Needs Tesseract (libtesseract >=3.04), Leptonica (>=1.71), Cython (build only), Pillow (optional), tessdata language files
  • Models: Tesseract traineddata, LSTM engine

What changed

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Data and scraping for AI

See all 69
Also in Data and scraping for AI
RankProjectScore
1mineruParses PDFs, images and Office files into Markdown and structured JSON Live demo ↗ (opens in a new tab)83.9k starscustom license91 out of 100
2doclingPython library that converts PDFs, Office files and more into structured text70.7k starsMIT91 out of 100
3opendataloader-pdfPDF parser that outputs Markdown, JSON with bounding boxes, and Tagged PDF30.3k starsApache-2.091 out of 100
#4agent-reachCLI that gives coding agents read access to login-gated and Chinese web platforms112.8k starsMIT88 out of 100
#5markitdownPython library and CLI that converts office files, PDFs and audio to Markdown196.2k starsMIT87 out of 100