Deepgram
A voice API platform for transcribing speech, generating audio and building real-time conversational agents.
Audio
Developer Tools
API available
AI-native
Pricing
Free trial
Paid, with a limited free trial.
Category
Audio & Voice
Platforms
Web, API
Languages
Multilingual
Last updated
July 18, 2026
Overview
Deepgram provides models and APIs for speech recognition, text-to-speech and conversational voice agents. The platform processes live streams and recorded files, with features such as speaker detection, formatting and language identification.
It is primarily designed for developers and businesses embedding voice into a product. A web console and initial credits support evaluation, after which billing depends on usage, selected models and enabled features.
It is primarily designed for developers and businesses embedding voice into a product. A web console and initial credits support evaluation, after which billing depends on usage, selected models and enabled features.
Key features
- Real-time and prerecorded transcription with language and speaker detection
- Text-to-speech for generating audio responses from an application
- Voice Agent API combining several components of spoken interaction
Best use cases
- Transcribe calls, meetings, media files or live audio streams
- Build a voice agent for support, scheduling or interactive services
- Add captions, audio search or speech analytics to a product
Pros
- Architecture designed for integrations and real-time processing
- Support for many languages in speech recognition workflows
- Usage-based pricing with initial credits for API evaluation
