mindee/doctr
docTR (Document Text Recognition) - a seamless, high-performing & accessible library for OCR-related tasks powered by Deep Learning.
Implements a two-stage pipeline combining deep learning models for text detection (localizing words via DBNet) and recognition (identifying characters via CRNN), with support for rotated documents and multi-class extraction through the KIE predictor. Processes PDFs, images, and webpages with flexible bounding box handling and exports results as nested document structures (Page/Block/Line/Word) or JSON. Built on PyTorch with optional visualization and HTML synthesis capabilities.
5,956 stars. Actively maintained with 2 commits in the last 30 days.
Stars
5,956
Forks
627
Language
Python
License
Apache-2.0
Category
Last pushed
Mar 09, 2026
Commits (30d)
2
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/mindee/doctr"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Compare
Related frameworks
breezedeus/CnSTD
CnSTD: 基于 PyTorch/MXNet 的 中文/英文 场景文字检测（Scene Text Detection）、数学公式检测（Mathematical Formula...
felixdittrich92/OnnxTR
OnnxTR a docTR (Document Text Recognition) library Onnx pipeline wrapper - for seamless,...
JaidedAI/EasyOCR
Ready-to-use OCR with 80+ supported languages and all popular writing scripts including Latin,...
githubharald/SimpleHTR
Handwritten Text Recognition (HTR) system implemented with TensorFlow.
bgshih/crnn
Convolutional Recurrent Neural Network (CRNN) for image-based sequence recognition.