docTR

docTR is an open source Python library for optical character recognition and document text parsing.

docTR is a Python library for optical character recognition powered by PyTorch. It provides tools to parse textual information, detect layouts, and recognize text from PDF files and images. The project uses a two-stage approach involving text detection and text recognition architectures. It is designed for developers who need to integrate text extraction and document analysis capabilities into their applications. The library supports various model architectures and includes features for handling rotated pages, table structure recognition, and key information extraction.

docTR at a glance

Best for: Software engineer
Runs on: Python, Linux, macOS, Windows, Docker, open source
Pricing: Free
Released: January 2021
Category: AI productivity tool
Website: github.com/mindee/doctr

Key features

Text detection
Text recognition
Layout analysis
Table structure recognition
Key information extraction
PDF and image support

Use it for

Extracting text from PDF documents
Detecting document layouts and regions
Recognizing text in images
Parsing key information from documents

Alternatives to docTR

deepdoctection — Orchestrates layout analysis, OCR, and model fine-tuning in PyTorch pipelines.
Docling — Provides advanced PDF understanding and local execution for sensitive data.
RapidOCR — It converts PaddleOCR models to ONNX format for fast inference and deployment.
MinerU — Parses complex documents and tables into structured markdown for agents.
PDF Inspector — Detects if a PDF is scanned or text-based to avoid unnecessary OCR costs.
MiniCPM OCR — Efficient document parsing with correctness assessment and high throughput.
DocETL — Extracts structured data and links knowledge using large language models.
OpenDataLoader PDF — It ranks first in benchmarks for layout and table extraction without needing a GPU.
MarkItDown — Converts documents into LLM-friendly Markdown for better token efficiency.
PaddleOCR — Open-source OCR and document parsing for 100+ languages
NuMind — Achieves high accuracy in document extraction and outperforms generalist models.
Sparrow — Extracts structured data locally from documents without cloud dependencies.

All docTR alternatives