Deepgram integration: real-time speech-to-text for voice agents
Deepgram powers real-time transcription for StrideOps.ai voice agents - the default STT, including the Flux model with built-in turn detection.
- Connects with
- API key (bring your own)
- What syncs
- Speech-to-text layer
Speech-to-Text (STT)
What you can do with Deepgram
Deepgram is the default speech-to-text engine for StrideOps.ai voice agents - one of six STT providers you can choose per agent. It streams transcription in real time so the agent understands callers instantly, and the Flux model adds built-in conversational turn detection.
- Real-time streamed speech-to-text (nova-3)
- Flux model with built-in turn detection
- Low-latency transcription for natural back-and-forth
- A searchable transcript for every call
- Swappable with other STT providers per agent
Setup
How it works
Add your Deepgram API key (or use the default)
Select Deepgram as the agent's STT provider
Optionally enable the Flux model for built-in turn detection
What teams use it for
- Accurate, low-latency understanding on live calls
- Searchable transcripts for every conversation
- Tighter turn-taking with Flux turn detection
FAQ
Questions, answered
What makes Deepgram a good fit for voice agents?
Real-time streaming transcription keeps latency low, and the Flux model adds built-in turn detection for more natural conversations.