RapidAI/RapidOCR

📄 Awesome OCR multiple programing languages toolkits based on ONNX Runtime, OpenVINO, MNN, PaddlePaddle, TensorRT and PyTorch.

What it solves

RapidOCR provides a fast, open-source Optical Character Recognition (OCR) tool designed for rapid offline deployment across multiple platforms and languages. It addresses the engineering overhead and deployment complexity often associated with deep learning OCR models, specifically by optimizing the inference process for terminal devices.

How it works

The project converts models from PaddleOCR into the highly compatible ONNX format. This transition simplifies and accelerates inference deployment on various devices. The system supports multiple programming languages (Python, C++, Java, and C#) and can be run using various inference engines such as ONNX Runtime, TensorRT, OpenVINO, and PyTorch.

Who it’s for

Developers who need to integrate high-performance, offline OCR capabilities into their applications without the heavy resource requirements of full training frameworks.

Highlights

  • Multi-platform and Multi-language: Supports Linux, Windows, and Mac, with default support for Chinese and English recognition.
  • Extreme Speed: Focused on low resource consumption and fast response times.
  • Extensive Compatibility: Works across multiple languages and inference engines.
  • Customizable: Users can fine-tune models using PaddleOCR and then deploy them via RapidOCR.

Related

  • Project
    PaddlePaddle/PaddleOCR

    PaddleOCR is an open‑source OCR and document‑AI toolkit that converts images/PDFs into structured, LLM‑ready JSON or Markdown. It offers multilingual scene‑text models (PP‑OCRv6), a lightweight vision‑language parser (PaddleOCR‑VL‑1.6) for tables, formulas, and ancient scripts, and a structure‑aware pipeline (PP‑StructureV3). Models run on CPU, GPU, XPU, NPU or via ONNX/TensorRT, and integrate with RAG/agent platforms like Dify and RAGFlow.

  • Project
    microsoft/onnxruntime

    A cross-platform machine learning accelerator that optimizes inference and training for models from various frameworks like PyTorch and TensorFlow.

  • Project
    zibo-chen/rust-paddle-ocr

    A lightweight Rust OCR library based on PaddleOCR models and the MNN inference runtime, providing text detection and recognition across multiple languages.

  • Dispatch
    PP-OCRv6 release: 50-Language OCR from 1.5M to 34.5M Parameters

    PaddlePaddle has released PP-OCRv6, a scalable OCR model family supporting 50 languages with parameter counts ranging from 1.5M to 34.5M.

  • Dispatch
    PaddleOCR 3.5 release notes / what's new

    PaddleOCR 3.5 introduces Hugging Face Transformers as a supported inference backend, allowing OCR and document parsing models to integrate more seamlessly into PyTorch-based workflows.