Skip to main content
Telnyx hosts Deepgram models on dedicated GPUs. Access them through the Telnyx plugin: 
from livekit.plugins import telnyx

stt = telnyx.deepgram.STT(model="nova-3", language="en")
The plugin provides two STT classes:
  • telnyx.deepgram.STT — Recommended. Connects to Deepgram models hosted on Telnyx GPUs. Takes a model parameter (nova-3, nova-2, flux).
  • telnyx.STT — Connects to Telnyx’s own transcription engine (default) or Deepgram via transcription_engine="Deepgram". Takes a transcription_engine parameter instead of model.
For most use cases, telnyx.deepgram.STT is the simpler interface.
Plugin source on GitHub

​
Available models

Latest generation, best accuracy. 
stt = telnyx.deepgram.STT(
    model="nova-3",
    language="en",
    interim_results=True,
    keyterm=["YourBrand", "custom-term"],  # keyword boosting
)

​
Nova-2

Previous generation, stable and reliable. Uses weighted keyword boosting. 
stt = telnyx.deepgram.STT(
    model="nova-2",
    language="en",
    interim_results=True,
    keywords=["YourBrand:2.0", "custom-term:1.5"],
)

​
Flux

Experimental, with built-in end-of-turn detection. Designed for real-time voice agents. 
stt = telnyx.deepgram.STT(
    model="flux",
    language="en",
    interim_results=True,
    keyterm=["YourBrand", "custom-term"],
    eot_threshold=0.5,
    eot_timeout_ms=3000,
    eager_eot_threshold=0.3,
)

​
Parameters

ParameterDefaultDescription
modelnova-3Model to use (nova-3, nova-2, flux)
languageenLanguage code
interim_resultsTrueStream partial transcriptions
keyterm—Keyword boosting (Nova-3, Flux)
keywords—Weighted keyword boosting (Nova-2)
eot_threshold—End-of-turn confidence threshold (Flux only)
eot_timeout_ms—End-of-turn timeout in ms (Flux only)
eager_eot_threshold—Eager end-of-turn threshold (Flux only)