Fast English support lines
nova-3 is the lowest-latency option for English, which keeps dead air short on inbound and outbound calls. The same config works for both.
Deepgram turns the caller's speech into text as they talk, so the LLM can understand it and reply.
Telephony
Phone network
STT · You are here
Listener
LLM
Reasoning
TTS
Voice
Tools
Actions
Connectors for this stage
Transcribers (Speech-to-Text)
| Model | Languages | Best for |
|---|---|---|
nova-3 | English | Lowest latency, recommended default |
nova-3-medical | English | Clinical terminology |
nova-2 | 30+ languages | Multilingual agents |
nova-2-phonecall | English | Phone audio with background noise |
nova-2-conversationalai | English | AI agent conversations |
nova-2-meeting | English | Multiple speakers |
nova-2-medical | English | Medical terminology |
nova-2-finance | English | Financial and trading terminology |
nova-2-atc | English | Air traffic control |
nova-2-drivethru | English | Quick-service ordering |
nova-2-automotive | English | In-car applications |
| Flux (English) | English | Turn detection inside the model |
| Flux (Multilingual) | Multiple, with per-turn language identification | Multilingual agents without a preset language |
For endpointing, encoding and every model, see the Deepgram setup guide in the Bolna docs.
For Indian languages, the docs note Sarvam may give better accuracy.
nova-3 is the lowest-latency option for English, which keeps dead air short on inbound and outbound calls. The same config works for both.
Use nova-3-medical, nova-2-medical or nova-2-finance so clinical or trading terms come through correctly.
Flux can start the LLM response before the caller fully stops and detects barge-ins without VAD delay. See the Deepgram Flux guide.
The docs recommend Deepgram nova-2 for high-volume calling where cost matters, and it sits inside Bolna's flat per-minute rate. See pricing.
ElevenLabs Scribe v2 Realtime transcribes callers for a Bolna voice agent in about 150ms, across 90 languages, with word-level timestamps and auto language detection.
Google's Gemini 3.5 Transcribe gives a Bolna voice agent streaming speech-to-text in 19 language codes, including Hindi, Tamil and Bengali, on phone and web calls.
Gladia's Solaria model gives a Bolna voice agent real-time speech-to-text in 100+ languages with code-switching, custom vocabulary and under 300ms of latency.
Understands callers and decides the reply.
18 providersConnects agents to the phone network.
14 providersTurns caller speech into text in real time.
10 providersTurns replies into natural speech.
12 providersLets agents act during and after calls.
7 providersWalk through provider choice, stack trade-offs and the exact workflow you want Bolna to automate.
Secure, compliant and dependable, from the first pilot to full scale.
Built for your busiest hour. Even if the model fails, the call doesn’t
A dedicated team, with service levels shaped around your business
Data residency, encryption, access controls, full audit trails and more
Run Bolna in your own cloud or on-premise, your data remains yours
Certified by
