As reported by YourStory, Indian artificial intelligence (AI) startup Sarvam AI officially launched its innovative speech recognition system, Sarvam Audio, in February 2026. Specifically engineered to navigate India's unique linguistic landscape—characterized by multilingualism, frequent code-mixing (the blending of multiple languages in speech), and diverse writing systems—the system integrates advanced features such as contextual awareness, speaker differentiation, and five distinct output modes. These modes include standardized text output, Romanized transcriptions, intelligent translations, and more, catering to a wide range of user needs.

Sarvam Audio is particularly well-suited for deployment in customer service, logistics, finance, and other sectors where effective communication across languages is critical. To ensure high accuracy, even in noisy environments, the system intelligently leverages conversation history to enhance recognition rates. Additionally, it offers developers the flexibility to customize output formats according to specific application requirements. Currently, Sarvam AI has made an online experience platform available, allowing users to explore the capabilities of Sarvam Audio firsthand.