Voce: l’interfaccia utente originale.

Voxtral is a new generation of open speech understanding models designed to move beyond simple transcription and provide deep semantic comprehension of spoken language. It bridges the gap between limited open-source options and expensive proprietary APIs by providing state-of-the-art accuracy and native multilingual support in a flexible package. The models can process long audio files up to 30 minutes, perform Q&A directly on voice content, and even execute system commands through function calling. Released under an open-source license, Voxtral is available in different sizes for production or local use and can be accessed via an API at less than half the price of competing services.