Deepgram
APIs for recorded and streaming transcription, speech generation, and voice agents.
- Recorded and live speech recognition
- Text-to-speech and voice-agent APIs
- Usage rates depend on the model and speech task
Deepgram provides speech APIs for applications that handle recordings or live conversations. Speech recognition, speech generation, and voice-agent products have different endpoints and runtime requirements.
Recorded or live audio
Recorded transcription suits files that can be submitted as a completed job. Streaming recognition processes audio as it arrives and is useful for captions or voice interactions. Supported languages and options depend on the model and endpoint.
How pricing works
Transcription is generally priced by audio duration, with rates that depend on the model and features. Speech generation and voice agents use their own billing units. Pay-as-you-go access, committed plans, and promotional rates are different arrangements. Compare a recurring workload using the intended rate, not only an introductory credit. For another hosted transcription option, see AssemblyAI .
Sources
- Tags:
- Voice
- Developer Tools