Introducing GPT-transcribe and GPT-live-transcribe in Microsoft Foundry

A transcription model hears “account number 8-4-7-2” but returns “account number eighty-four seventy-two.” A single error can break a downstream automation workflow. Developers building voice applications need transcription models that can handle real-world audio conditions, natural speech patterns, and business-critical details, including codes, dates, addresses, account numbers, mixed-language conversations, specialized terminology, and quiet or low-volume speech. GPT-transcribe and GPT-live-transcribe do just that and are available in Microsoft Foundry today .