---
type: "post"
slug: "deepgram-flux-for-real-time-voice-agents"
title: "Deepgram Flux for real-time voice agents"
description: "Deepgram's Flux TTS treats speech synthesis as a conversation rather than a series of narration requests, because per-turn requests cannot hold tone across a live call."
retrieval_nugget: "Deepgram's Flux TTS treats speech synthesis as a conversation rather than a series of narration requests, because per-turn requests cannot hold tone across a live call."
published_at: "2026-08-31"
updated_at: "2026-09-12"
record_date: "2026-08-25"
date_kind: "discovered_at"
topics: ["agents","ai","voice-agents"]
entities: ["Deepgram"]
source_url: "https://deepgram.com/learn/introducing-flux-tts-conversation-native-text-to-speech-for-real-time-voice-agents"
source_title: "Introducing Flux TTS: Conversation-Native Text to Speech for Real Time Voice Agents"
source_domain: "deepgram.com"
source_terms: ["Deepgram","Flux","real-time","voice","model"]
summary_word_count: 192
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:deepgram-flux-for-real-time-voice-agents"
status: "published"
source_urls: ["https://deepgram.com/learn/introducing-flux-tts-conversation-native-text-to-speech-for-real-time-voice-agents"]
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:70ff88c2af434262a68b832e9f85096118cea0eb11d3f96e4a691bb9ce6e515c","reviewed_at":"2026-09-12T10:00:00Z","source_evidence_count":1,"verified_claim_count":2,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":2,"implication_count":1,"watch_condition_count":1,"related_record_count":2}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Deepgram launched Flux TTS, the second model in its Flux family, describing it as the first conversation-native text-to-speech model built for real-time voice agents; it follows Flux STT, which reworked speech recognition for live conversations, and is free until 12 September.","observed_facts":[{"text":"Flux TTS is the second model in the Flux family and is presented as conversation-native text-to-speech for real-time voice agents.","source_urls":["https://deepgram.com/learn/introducing-flux-tts-conversation-native-text-to-speech-for-real-time-voice-agents"]},{"text":"Deepgram states that existing text-to-speech was built for narration such as audiobooks, voiceover and IVR prompts.","source_urls":["https://deepgram.com/learn/introducing-flux-tts-conversation-native-text-to-speech-for-real-time-voice-agents"]}],"mechanism":"The first is tone: because each line is a fresh request and the model does not know it is inside a conversation, an empathetic tone cannot hold across a long call and a shift in the customer's mood cannot carry over.","why_now":"Today's text-to-speech was built for narration — audiobooks, voiceover, IVR prompts — and Deepgram's argument is that dropping a narration model into a live conversation breaks in identifiable ways.","implications":["The failure is architectural rather than a quality problem, which is why tuning a voice does not fix a call that drifts by minute three."],"evidence_boundary":"This is a launch post: it names the failure modes and the promotional window without publishing comparative latency or listener evaluations.","watch_conditions":["The condition to watch is whether Deepgram publishes turn-level measurements against a narration-trained baseline, because conversation-native is a claim about state handling that only a multi-turn test can show."],"related_records":[{"url":"https://newruntime.com/posts/gpt-live-continuous-media-loop","relation":"Prior coverage of separating the conversation loop from deep reasoning."},{"url":"https://newruntime.com/topics/agents","relation":"Topic hub for latency-shaped agent design."}]}
routes: {"html":"https://newruntime.com/posts/deepgram-flux-for-real-time-voice-agents/","markdown":"https://newruntime.com/posts/deepgram-flux-for-real-time-voice-agents.md","json":"https://newruntime.com/posts/deepgram-flux-for-real-time-voice-agents.json"}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
---

# Deepgram Flux for real-time voice agents

## Retrieval answer

Deepgram's Flux TTS treats speech synthesis as a conversation rather than a series of narration requests, because per-turn requests cannot hold tone across a live call.

Deepgram launched Flux TTS, the second model in its Flux family, describing it as the first conversation-native text-to-speech model built for real-time voice agents; it follows Flux STT, which reworked speech recognition for live conversations, and is free until 12 September.

The diagnosis is more specific than the product claim. Today's text-to-speech was built for narration — audiobooks, voiceover, IVR prompts — and Deepgram's argument is that dropping a narration model into a live conversation breaks in identifiable ways. The first is tone: because each line is a fresh request and the model does not know it is inside a conversation, an empathetic tone cannot hold across a long call and a shift in the customer's mood cannot carry over. The failure is architectural rather than a quality problem, which is why tuning a voice does not fix a call that drifts by minute three.

Treating the turn as state rather than a request is the same separation described in [GPT-Live splitting the conversation loop from deep reasoning](https://newruntime.com/posts/gpt-live-continuous-media-loop/), and it belongs with the latency-shaped design decisions collected in the [agents hub](https://newruntime.com/topics/agents/).

This is a launch post: it names the failure modes and the promotional window without publishing comparative latency or listener evaluations. The condition to watch is whether Deepgram publishes turn-level measurements against a narration-trained baseline, because conversation-native is a claim about state handling that only a multi-turn test can show.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Agents](https://newruntime.com/topics/agents.md): Explore the agents topic hub.
- topic hub: [Ai](https://newruntime.com/topics/ai.md): Explore the ai topic hub.
- topic hub: [Voice Agents](https://newruntime.com/topics/voice-agents.md): Explore the voice-agents topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)