Umi-OCR is an offline optical character recognition tool that runs locally without requiring network connectivity.
The tool addresses the need for free, accessible text extraction from images and documents. It uses PaddleOCR as its recognition engine and provides a graphical interface built with Qt and QML. The software handles multiple input methods including screenshot capture, batch image import, and PDF document processing. It includes built-in language libraries for multilingual recognition and offers text post-processing capabilities to handle different document layouts and exclude unwanted regions like watermarks, headers, and footers. The tool also supports QR code scanning and generation.
Umi-OCR suits developers and end users who need reliable offline OCR without cloud dependencies or subscription costs. It works well for batch processing workflows, document digitization, and integration into other applications through its command-line interface and HTTP API. The software targets Windows 7 x64 and Linux x64 systems. Unlike cloud-based OCR services, this tool operates entirely offline and requires no network access after installation.
The project maintains active development with regular updates and multilingual support through community translation efforts. Bug reports and feature requests receive attention through the project's issue tracker. The codebase is organized to support both end-user applications and developer integration, with documentation provided for command-line usage and HTTP interface implementation. The project includes build instructions for developers working with Windows and Linux environments.
