To receive

SpeechEventType.INTERIM_TRANSCRIPT

events in LiveKit when using Deepgram STT with

interim_results=True

and

smart_format=True

, ensure the following: • Enable interim results: Confirm that

interim_results=True

is set when initializing the Deepgram STT instance. This parameter enables preliminary results before the final transcription is available. • Proper STT stream setup: Use a streaming STT instance, as interim transcripts are only available through the streaming interface. • Audio data flow: Continuously push audio frames to the STT stream using methods like

push_frame()

. Here is an example configuration in Python:

from livekit.plugins.deepgram import STT

stt = STT(
    interim_results=True,  # Required for interim transcripts
    smart_format=True,     # Enables smart formatting
    # other parameters...
)

The

smart_format=True

and

interim_results=True

parameters work together without conflicts. Both are passed to Deepgram's API in the live configuration. The

smart_format

parameter improves the readability of transcriptions and applies to both interim and final transcripts. If you are still not receiving interim transcript events, verify that: • Audio frames are being pushed to the STT stream continuously. • The transcript contains actual text content (interim transcripts are only emitted when there is text). Let me know if you need further assistance! Sources: livekit-plugins/livekit-plugins-openai/livekit/plugins/openai/stt.py | Deepgram integration guide | LiveKit Docs | agents/src/stt/stt.ts | SpeechEventType | LiveKit Agents