Speech Technology eWeekly - July 16, 2025

Speech Technology News

Deepgram Launches Saga, a Voice OS for Developers

Deepgram's Saga turns natural speech into cross-tool workflow execution via MCP.

Soul App Launches Full-Duplex Voice Model

With its Full-Duplex Voice Model, Soul App's AI gains autonomous speech timing control. 

Vonage Partners with AWS for AI Voice Agent Integration

The partnership pairs the Amazon Nova Sonic speech-to-speech foundation model with the Vonage Voice API.

Mistral Unveils Voxtral Open-Source AI Voice Model

Voxtral is an open-source audio model for high-quality speech recognition and semantic understanding.