A-Eye Web Chat Assistant is a client-side Chrome extension designed to help visually impaired users browse the web via voice and vision AI.
GeProVIs AI Screen Reader, the team's previous project, garnered significant praise but faced several limitations, including web security, cost, and privacy concerns.
A-Eye Web Chat Assistant is a more comprehensive solution that relies on Web AI and Google's Gemini for web browsing with superior accuracy and privacy-first security.
The extension provides real-time descriptions of images, conversation management, and dual AI models along with cross-platform compatibility and ease of use.
Users can chat with the extension with voice commands, while speech synthesis technology via Web Speech API read out responses.
The tool also uses AI models like Moondream1ForConditionalGeneration and RawImage from Transformers.js 3.0 for image processing and description.
The tool has been designed as a Chrome extension but the developers hope that it will be integrated into Chrome as a default tool as it is more beneficial and powerful.
The team is seeking an experienced developer to assist with a full refactoring of the codebase.
The project is based on Hong Kong Institute's IT diploma course and it has been developed by Vincent Wun and Li Yuen Yuen and is part of the GDG group.
The tool is 100% free and open source. It offers a lot of features such as real-time image description and text-to-speech for HTML element content.