Logo Lanfrica

aye-technology/fidel-scanner

Domain:

natural language processing

Record type:

softwaretools
Creator:
aye
Host:
Fidel Scanner is a privacy-focused web application that converts images and PDFs into editable text for Amharic, Tigrinya, Afaan Oromoo, and English. It uses AI (Tesseract.js) to process documents locally in your browser so no data ever leaves your device, featuring a side-by-side desktop interface for easy comparison # πŸ” Fidel Scanner **Fidel Scanner** is a professional, high-performance web application designed to convert images and PDF documents into editable text. Specifically optimized for Ethiopic scripts (**Amharic** and **Tigrinya**) and Latin scripts (**Afaan Oromoo** and **English**), it provides a seamless, high-speed OCR (Optical Character Recognition) experience directly in your browser. --- ## 🌟 Key Features * **Multi-Format Support:** Process standard image files (JPG, PNG, GIF) and multi-page PDF documents. * **Smart Page Selection:** For PDF files, users can select specific pages to scan, saving time and resources. * **Side-by-Side Dashboard:** Optimized desktop interface featuring two distinct cards for source viewing and result editing, allowing for easy real-time comparison. * **Local AI Processing:** Unlike most OCR tools, Fidel Scanner performs all AI computations locally on your device. Your sensitive documents never touch a server. * **Ethiopic Script Optimization:** Utilizes high-quality training models for accurate recognition of Ge'ez characters. * **Export Options:** Instantly copy extracted text to your clipboard or download it as a `.txt` file. --- ## 🌍 Supported Languages The application is specifically tuned for: * **Amharic (αŠ αˆ›αˆ­αŠ›)** * **Tigrinya (α‰΅αŒαˆ­αŠ›)** * **Afaan Oromoo (Qubee)** * **English** --- ## πŸ› οΈ Technical Stack Built with a modern, lightweight stack to ensure speed and portability: * **React (v18):** Core framework for state management and UI components. * **Tesseract.js (v5):** The engine responsible for Optical Character Recognition using deep-learning models. * **PDF.js:** Mozilla's library for rendering and manipulating PDF documents within the browser. * **Tailwind CSS:** For a responsive, "utility-first" design that adapts to all screen sizes. * **Phosphor Icons:** Minimalist and scalable icon set. * **Noto Sans Ethiopic:** Ensuring Ge'ez characters are rendered with maximum legibility. --- ## πŸ”’ Privacy & Security …