Key Features

What you can do

Multilingual Text Recognition

Supports recognition of text in 109 languages including digits, vertical text, and long text formats.

Pre-trained PPOCR Models

Includes lightweight and general models such as ppocr_mobile, ppocr_server, and ppocr_mobile_slim for various resource environments.

Document Parsing and Layout Analysis

Provides tools like PP-Structure and PaddleOCR-VL for analyzing complex layouts including tables, formulas, charts, and preserving structure in JSON/Markdown.

Information Extraction with PP-ChatOCRv4

Integrates ERNIE 4.5 for extracting information from documents to support advanced data processing.

Cross-Platform Installation and Deployment

Supports installation via PIP on Windows, Linux, and MacOS, with deployment options including self-hosted MCP server.