Overview

Enterprise-grade, streaming-first TTS. Aura-2 pairs sub-200ms latency with rock-solid reliability across thousands of concurrent sessions, on one unified STT+TTS stack. Transparent per-character pricing and flexible deployment. Built for contact centers and scale, not premium narration.

Features

Streaming First
Streaming-first model with sub-200ms latency for real-time agents.

Unified Voice stack
Shared infrastructure lowers latency and keeps pronunciation consistent

Mulitaccent Support
Localized accents, in 40+ english voices, tuned for professional, business contexts

Domain accuracy
Reliable pronunciation of healthcare, finance, and legal terminology

Flexible deployment
Public cloud, private cloud, and on-premises options

Transparent usage pricing
$0.030 per 1,000 characters with volume discounts

Pros and Cons

What Works Well
  • Enterprise-grade reliability and stability at scale
  • Very low latency (sub-200ms) for real-time agents
  • Unified STT+TTS pipeline reduces integration complexity
  • Flexible deployment including on-premises for compliance
WHAT could be better
  • Voices less refined/expressive than dedicated TTS leaders
  • No voice cloning
  • Developer-only, no consumer studio
  • TTS is a secondary focus behind Deepgram’s core STT

Pricing Plans

Pay as you go
Custom
Contact for Pricing
Growth (Annual)
$ 4000
per year

Customers

  • Contact centers running thousands of concurrent voice sessions
  • Developers building full STT+TTS voice pipelines
  • Enterprises in healthcare, finance, and legal needing reliability
  • High-volume voice-agent operators needing predictable cost

Company Information

Company Name
Deepgram
Company Founded Year
2015
Company Location
San Francisco, California, USA

Frequently Asked Questions

How is Aura text-to-speech priced?
Per 1,000 characters of input text: Aura-2 is $0.030 per 1K on Pay-As-You-Go ($0.027 on Growth), and Aura-1 is $0.0150 per 1K ($0.0135 on Growth).
Is there a free way to try it?
Yes. Deepgram gives $200 in free credit with no credit card required, usable across STT, TTS, Voice Agent, and Audio Intelligence. Credits don’t expire.
Is TTS billed by minute or by character?
Text-to-Speech (Aura) is billed by characters of input text, not by the duration of the output audio.
What’s the difference between Pay-As-You-Go and Growth?
Pay-As-You-Go has no minimums or expiration; Growth uses prepaid annual credits (starting around $4K/year) that save up to ~20% and raise concurrency limits.
Can I self-host or deploy on-premise?
Yes, self-hosted and private-cloud/on-prem deployment is available, handled through Enterprise agreements.
Does it support voice cloning?
No. It offers 40+ preset English voices with localized accents; it doesn’t provide voice cloning.
Is Deepgram HIPAA and SOC 2 compliant?
Yes. Deepgram is SOC 2 Type 1 & Type 2 certified, GDPR-ready with an EU data-residency endpoint, and signs BAAs for HIPAA on Enterprise.
What concurrency do I get for TTS?
Up to 45 concurrent requests (REST + WSS) on Pay-As-You-Go, and up to 60 on Growth.

Alternatives

inworld-ai-logo
Inworld
⭐️ 4.5
strength
Text-to-Speech Realism
Voice-first AI platform leading TTS realism, tuned for characters, games, and agents.
FROM
$ 25
per month
✅ Free Trial
View More

elevenlabs-symbol-logo1
Elevenlabs
⭐️ 4.7
strength
Voice cloning
The gold-standard AI voice platform for hyper-realistic text-to-speech and voice cloning.
FROM
$ 6
per month
✅ Free Trial
View More

cartesia-ai-logo
Cartesia
⭐️ 4.6
strength
Low latency TTS
Ultra-low-latency TTS API built on state space models for real-time voice agents.
FROM
$ 3
per month
✅ Free Trial
View More