NVIDIA has published a post on the Hugging Face blog introducing its latest transcription developments. The publication announces "NVIDIA Nemotron 3.5 Transcribe and Granary v2" as part of the company's speech technology portfolio. The post focuses on capabilities that allow operators to work across multiple spoken languages.
The announcement characterizes the system as "Intelligent Multilingual Transcription You Can Prompt and Adapt" within speech workflows. By coupling transcription with prompt-based adjustments, the tools aim to address varied linguistic environments. Granary v2 and Nemotron 3.5 Transcribe together define this multilingual transcription release.

