Nvidia Releases Free Speech-to-Text AI Tool for Recordings

Nvidia has released Nemotron 3.5 Transcribe, a free, downloadable AI tool that turns recordings into text. It has 30 billion settings and runs on one chip. It works in 40 languages, handles recordings up to one hour, and identifies up to eight speakers per file. It cannot process live streams. TechCrunch reported it on October 28, 2026.