I got this STT error today and the call choked after the greeting. Is there a way to add backup models with inference, or a recommended way to handle this for production ? `DEBUG livekit.agents using transcript io:

AgentSession

->

TranscriptSynchronizer

-> {"pid": 89222, "job_id": "AJ_7Tjv5Ct2Uh7J", "room_id":` ``RoomIO` "RM_JhoyCRvtMUCh"}`

ERROR  livekit.agents   Error in recv_task                                                     {"pid": 89222, "job_id": "AJ_7Tjv5Ct2Uh7J", "room_id":

Traceback (most recent call last):                                     "RM_JhoyCRvtMUCh"}

File

"/Users/matthew/AIAgents/livekit/hello-world-service-agent/.venv/lib/pyth

on3.13/site-packages/livekit/agents/utils/log.py", line 16, in

async_fn_logs

return await fn(*args, **kwargs)

^^^^^^^^^^^^^^^^^^^^^^^^^

File

"/Users/matthew/AIAgents/livekit/hello-world-service-agent/.venv/lib/pyth

on3.13/site-packages/livekit/agents/inference/stt.py", line 408, in

recv_task

raise APIError(f"LiveKit STT returned error: {msg.data}")

livekit.agents._exceptions.APIError: LiveKit STT returned error:

{"type":"error","session_id":"5d219d7e-b790-494d-8f5c-8962c57a09e6","d

ata":"Failed to connect to provider for session

5d219d7e-b790-494d-8f5c-8962c57a09e6: failed to connect to provider

nova-3: Provider deepgram error (status 400): websocket: bad handshake

is this error consistent or intermiteent?

Intermittent. Thanks I'll checkout the fallback adapters