Here’s an example of overriding the STT (Speech-to-Text) node in LiveKit:

from livekit import rtc
from livekit.agents import ModelSettings, stt, Agent
from typing import AsyncIterable, Optional

async def stt_node(
    self, audio: AsyncIterable[rtc.AudioFrame], model_settings: ModelSettings
) -> Optional[AsyncIterable[stt.SpeechEvent]]:
    async def filtered_audio():
        async for frame in audio:
            # insert custom audio preprocessing here
            yield frame
    
    async for event in Agent.default.stt_node(self, filtered_audio(), model_settings):
        # insert custom text postprocessing here
        yield event

This example demonstrates how to add a noise filtering step to the STT process. The method first creates a filtered audio stream with custom preprocessing, then passes it to the default STT node implementation, and finally allows for postprocessing of the transcribed text before yielding the speech events. Sources: Voice pipeline nodes | LiveKit Docs