Be first!
Deepgram
Enterprise-grade Voice AI APIs for real-time speech-to-text, text-to-speech, and conversational voice agents.
FreemiumAI Agents
Deepgram
Enterprise-grade Voice AI APIs for real-time speech-to-text, text-to-speech, and conversational voice agents.
Freemium
Categories
Overview
Deepgram offers advanced voice AI solutions including speech-to-text, text-to-speech, and a unified Voice Agent API that integrates conversational AI with real-time transcription and natural voice synthesis. It supports over 36 languages with ultra-low latency, high accuracy, and customizable models tailored for industries like healthcare, customer support, and media. Trusted by enterprises and startups, Deepgram enables scalable, secure, and cost-effective voice AI experiences through flexible cloud and self-hosted deployments.
Featured Video
Video via YouTubeWatch on YouTube
Key Features
1
Unified Voice Agent API
Combines speech-to-text, text-to-speech, and large language model orchestration into a single API to reduce complexity, latency, and cost for building conversational AI agents.
2
Flux Conversational STT Model
A speech-to-text model optimized for real-time conversation with built-in turn detection, natural interruption handling, and sub-300ms latency for human-like voice agents.
3
Nova-3 Transcription Model
High-performance speech-to-text model offering top accuracy, multilingual support, and noise robustness for production transcription needs.
4
Industry-Tuned and Custom Models
Specialized models optimized for domains like healthcare, legal, and finance, plus custom models trained on proprietary datasets for maximum accuracy.
5
Audio Intelligence Features
Includes summarization, topic detection, sentiment analysis, and intent recognition powered by task-specific language models that work with or without transcription.
6
Multichannel Audio Support
Ability to transcribe multichannel audio with speaker diarization and separate channel billing for accurate transcription in overlapping speech scenarios.
7
Text-to-Speech with Natural Voices
Responsive, natural-sounding text-to-speech models designed for high-throughput voicebots and conversational AI applications, billed per character.
8
Enterprise-Grade Security and Scalability
Offers cloud and self-hosted deployment options, priority support, and compliance-ready solutions for large volume and sensitive data environments.
Who It's For
Audience 1#1
Customer Support Transcription
Accurately transcribe and analyze customer calls in real-time to improve support quality and agent performance.
Audience 2#2
Healthcare Documentation
Enable HIPAA-compliant medical transcription with specialized vocabulary and real-time clinical workflow support.
Audience 3#3
Conversational AI Agents
Build voice agents that listen, understand, and respond naturally using integrated speech-to-text, LLMs, and text-to-speech.
Audience 4#4
Media Captioning and SEO
Generate accurate captions and transcripts for podcasts, videos, and broadcasts to enhance accessibility and searchability.
Audience 5#5
Speech Analytics and Insights
Extract sentiment, intent, and topics from conversations to drive actionable business intelligence.
How to Use
1
Sign Up and Get API Key
Create a free Deepgram account to access your API key and start using the platform with $200 in free credits.
2
Choose Your Speech-to-Text Model
Select from Flux for real-time conversation, Nova-3 for transcription accuracy, or industry/custom models based on your needs.
3
Integrate Audio Input
Stream live audio or upload pre-recorded files to the Deepgram API for transcription and analysis.
4
Configure Features
Enable optional features like speaker diarization, keyterm boosting, redaction, and smart formatting to tailor output.
5
Use Voice Agent API for Conversational AI
Leverage the unified API to build voice agents that combine STT, LLM orchestration, and TTS for natural interactions.
Pricing
Pricing details are gathered from the official Deepgram website and are provided for reference only. Always confirm the latest information directly with the vendor.
| Plan | Price | Highlights |
|---|---|---|
| Pay As You Go | Free $200 credit then pay-as-you-go | Access all speech-to-text, text-to-speech, and audio intelligence endpoints
|
| Growth | From $4,000 | All Pay As You Go features
|
| Enterprise | Custom Pricing | Custom-trained speech-to-text models
|
Found a change in pricing? We welcome corrections. Reach out so we can keep this listing accurate.
Pros & Cons
Pros
- Unified API simplifies building conversational AI agents by integrating STT, TTS, and LLM orchestration.
- Ultra-low latency transcription with sub-300ms delay supports real-time applications.
- Supports over 36 languages and dialects for global reach.
- Custom and industry-tuned models improve accuracy for specialized domains.
- Flexible pricing plans including pay-as-you-go and enterprise options with self-hosting available.
Cons
- Pricing can be complex due to multiple models and add-ons like redaction and keyterm prompting.
- Some advanced features require contacting sales, limiting transparency for smaller users.
- Text-to-speech currently supports only English language.
- Voice Agent API pricing depends on WebSocket connection time, which may be harder to estimate.
- Limited public documentation on detailed LLM integration options and tiers.
Frequently Asked Questions
What is included in the Voice Agent API pricing?
Voice Agent API pricing is based on WebSocket connection time and includes usage of Deepgram's speech-to-text, text-to-speech, and built-in LLM support. Using your own LLM may reduce costs.
How is multichannel audio billed?
Each audio channel is transcribed and billed separately. If multichannel is not enabled, audio is converted to mono and billed as a single channel.
Which speech-to-text models does Deepgram offer?
Deepgram offers Flux for real-time conversation, Nova-3 for high accuracy transcription, Industry-Tuned models for specialized domains, and Custom models trained on proprietary data.
Can Deepgram transcribe live streaming audio?
Yes, Deepgram supports live streaming audio transcription with latency under 300 milliseconds for real-time applications.
What languages are supported?
Deepgram supports transcription in over 36 languages and dialects, and English for text-to-speech.
