Official announcements from major browser developers (e.g., Google Chrome, Mozilla Firefox, Microsoft Edge)

OpenAI Introduces Whisper WebGPU for Real-Time In-Browser Speech Recognition

Jun 7, 2024, 04:38 PM

OpenAI has achieved a significant milestone with the introduction of real-time in-browser speech recognition using its Whisper model. This breakthrough, known as Whisper WebGPU, allows for speech-to-text functionality that operates entirely on-device within a web browser. The technology leverages Transformers.js and ONNX Runtime Web to provide multilingual transcription across 100 different languages. Developed by a Hugging Face Engineer, known by the nickname ‘Xenova’, this development is seen as a game changer, particularly due to its ability to run locally without the need for cloud-based processing.