Deepgram launches Flux TTS: Real-time streaming voice with sub-200ms latency
CodeByPoonam · x · 2026-08-17
Deepgram has officially launched Flux TTS, a text-to-speech model purpose-built for real-time conversations.
Key Features:
- Streaming Generation: Unlike traditional TTS that waits for full text generation, Flux starts speaking as tokens arrive.
- Ultra-low Latency: Sub-200ms response time enables truly natural conversational flow.
- Conversation-Aware: The model reads the entire conversation context, maintaining emotional and tonal consistency, and handles interruptions and conversational cues.
Architecture: Flux STT (listening) and Flux TTS (speaking) run as a single integrated stack, with Deepgram orchestrating the timing between listening and speaking. This product is ideal for AI agents, call centers, and meeting assistants.
More from Multimodal
- MLP-Splatting (ECCV 2026) code released: MLPs as light-field primitives instead of Gaussians — AjdDavison · 2026-10-16
- Digital human Fae codes her own facial expressions frame-by-frame with Claude Opus 5.5 and FACS — repligate · 2026-10-16
- Meta's Omnilingual ASR, covering 1,600 languages, named a TIME Best Invention of 2026 — AIatMeta · 2026-10-16
- ATAS: Zhejiang-led team builds world-state memory framework for consistent agentic video generation — CSProfKGD · 2026-10-16
- Flux 3 Klein's Default Templates Are Broken Without Its Prompt Upscaler — StrictMove5839 · 2026-10-16
- Kroma Touted as the Most Well-Rounded Image Model Yet on Civitai — Different_Fix_2217 · 2026-10-16