MAI-Transcribe-1

Whisper is an automatic speech recognition (ASR) system developed by OpenAI for converting spoken language into text. It is trained on 680,000 hours of multilingual and multitask audio data collected from the web. The model is designed to handle diverse accents, background noise, and technical language with high accuracy. Whisper supports transcription in multiple languages as well as translation into English. It uses an encoder-decoder Transformer architecture to process audio inputs and generate text outputs. The system can also perform tasks like language identification and timestamp generation. Overall, Whisper enables developers to build robust voice-enabled applications with ease.

MAI-Transcribe-2 is Microsoft AI’s most capable transcription model yet, designed to deliver fast, accurate speech recognition across a broad range of real-world audio. It supports speaker diarization to distinguish speakers and attribute words to the right person, along with word-level timestamps for precise alignment, search, navigation, and editing. Keyword biasing helps recognize domain-specific terminology, abbreviations, names, and other terms that can be difficult to distinguish from context alone. Developers can choose between configurable transcription styles: a verbatim setting that preserves filler words and false starts for compliance and analysis, or a clean setting that removes fillers for more readable captions, notes, and published transcripts. The model supports code-switching for conversations that naturally move between languages, including blended language pairs such as Hinglish and Spanglish, and can automatically identify the language being spoken.