Bumblebee Companion

Client: British Wireless for the Blind Fund

Bumblebee Companion is a groundbreaking initiative designed to enhance accessibility for individuals with visual impairments. The project integrates the power of AI, specifically GPT-4o, to create an intelligent, real-time assistant that simplifies the process of accessing and selecting radio stations. This assistant is built within the larger Companion Framework and aims to empower users with disabilities by providing a seamless and personalized listening experience.

Key Features:

Real-Time Radio Selection: Bumblebee Companion helps users navigate a vast library of over 16,000 radio stations. It allows users to select stations based on personal preferences, including language, schedule, and genre, making the process more intuitive and accessible.

Function Calling Capabilities: The assistant leverages function calling to provide recommendations and help users quickly tune into the stations that match their preferences. By pulling information such as live broadcasts, program schedules, and station attributes, Bumblebee Companion ensures a tailored experience for each user.

Personalized Interaction: With the use of machine learning models fine-tuned to individual preferences, Bumblebee Companion learns from user interactions to provide better recommendations over time. This creates a dynamic and evolving user experience, adapting to changes in listening habits and preferences.

Multi-Modal Integration: The project is designed to work across various input modes, including voice commands and text inputs, ensuring ease of use for individuals with different levels of impairment.

Support for Visual Disabilities: Bumblebee Companion is fully optimized for users who rely on assistive technologies. Its interface is built with accessibility in mind, ensuring compatibility with screen readers and other tools commonly used by individuals with visual impairments.

Technical Stack:

AI & LLM Integration: Powered by GPT-4o, Bumblebee Companion utilizes natural language processing (NLP) to understand user preferences and provide intelligent recommendations.

Data Sources: A comprehensive dataset of radio stations, including metadata like language, broadcast times, and genres, forms the core of the app's recommendation system.

Real-Time API: Bumblebee Companion uses the OpenAI Realtime API for fast, efficient, and accurate responses to user queries.

Mobile & Web Support: The platform is accessible via mobile applications, providing on-the-go support for users, and it can be expanded for other platforms, ensuring versatility and ease of use.

Bumblebee Companion is not just a technical solution; it’s a transformative tool designed to bring greater autonomy and enjoyment to individuals with visual impairments, offering them the opportunity to engage with the world of radio on their own terms.