Deepgram Text-to-Speech API

What it does

Deepgram Text-to-Speech API turns written text into spoken audio. Developers send text, either as a single request or as a continuous text stream, and receive streamed audio back at 24kHz. The product is built around real-time voice agents, where tone and pacing are meant to carry across conversational turns.

Its distinguishing features are conversational context and interruption handling. Flux TTS is described as reading the whole conversation rather than only the current line, and the server reports what was heard when a caller interrupts. Voices come in English plus several other languages, and Aura-2 and Aura-1 are offered as lower-cost, high-volume models.

This suits developers building voice agents, IVR, or support flows that need low latency and predictable per-character billing. Access requires an API key from the Deepgram console; a free credit is offered to start. Teams needing strict data residency can use self-hosted or private cloud deployment, which is listed for Enterprise customers.

What it can do

• Stream generated speech from text
• Maintain conversational tone and pacing with supported models
• Report interruption progress for conversational applications
• Generate speech in supported languages and voices
• Deploy through the hosted API or an agreed enterprise environment

Works with

• Twilio
• LiveKit
• Daily
• Pipecat
• jambonz

Example workflows

Real-time voice agent reply

Illustrative example. Input: a text response from an LLM for a support caller. Steps: send the text to the TTS API, stream the returned audio to the caller, and if the caller interrupts, use the reported heard-content to resume. Output: streamed spoken reply that begins within a short delay.

Batch-style text stream narration

Illustrative example. Input: a continuous stream of text segments. Steps: send the segments through the continuous text stream endpoint and collect the returned audio chunks. Output: a sequence of audio chunks ready for playback or storage.

Before you choose

• Flux TTS is priced per 1,000 characters, so long scripts and repeated prompts increase cost directly
• Concurrency limits apply per project and differ by plan and region
• Self-hosted deployment is listed for Enterprise and requires NVIDIA GPUs per the FAQ

Listed under

• Support
• Voice and phone agents
• Conversational AI
• Customer support
• Developers
• Enterprise
• Healthcare
• Low latency

Get started

New accounts receive a $200 credit with no credit card required, per the pricing page.

1. Create a Deepgram console account at console.deepgram.com/signup and generate an API key
2. Install the Python, JS, Go, .NET or Java SDK, or call the REST API directly
3. Select a voice and language, then send a test text request from the Playground or API reference

Billed per 1,000 characters of input text on pay-as-you-go. A $200 credit is offered to new accounts and does not expire. The Flux TTS promotion of $1 credit for every $1 spent, up to $500, runs through Dec. 31, 2026. Check the pricing page for current rates.

Source: deepgram.com/pricing, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• Context-aware text-to-speech API for voice agents | Deepgram ↗deepgram.com
• Deepgram pricing | Scalable speech-to-text, text-to-speech & Voice Agent APIs | Deepgram ↗deepgram.com
• Welcome to Deepgram's docs | Deepgram docs ↗developers.deepgram.com

Company website

deepgram.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.