Deepgram speech-to-text and voice models now available natively on Together AI
Action Required
Developers can now build more robust and responsive real-time voice agents by leveraging Deepgram's advanced speech models within the Together AI platform, improving user experiences and operational efficiency.
AI Impact Summary
Deepgram is integrating its speech-to-text and voice models natively into Together AI's Dedicated Model Inference platform. This allows developers to leverage Deepgram's purpose-built STT and TTS models – including conversational STT (Flux), production transcription (Nova-3), and enterprise TTS (Aura-2) – within the Together AI ecosystem, simplifying the development of real-time voice agents and reducing latency. This capability expands Together AI's offerings and provides a more streamlined solution for voice-based applications.
Affected Systems
- Date
- 2 Apr 2026
- Change type
- capability
- Severity
- high