The error

ValueError: cannot construct SpeechConfig with both region and endpoint or host information

appears when you configure Azure STT in LiveKit with conflicting authentication parameters. The Azure Speech SDK requires you to supply exactly one method of authentication at a time—providing both

speech_region

and

speech_endpoint

or

speech_host

will cause this error. To resolve this, ensure your Azure STT configuration matches only one of these patterns: 1. Option 1: Set

speech_host

only (do not set

speech_region

or

speech_endpoint

). 2. Option 2: Set both

speech_key

and

speech_region

(do not set

speech_host

or

speech_endpoint

). 3. Option 3: Set both

speech_auth_token

and

speech_region

(do not set

speech_host

or

speech_endpoint

). 4. Option 4: Set both

speech_key

and

speech_endpoint

(do not set

speech_region

or

speech_host

). Double check your configuration and remove any parameters that are not part of a valid combination above. Environment variables may also be used, but avoid setting overlapping variables for region and endpoint/host simultaneously. More details can be found in the LiveKit documentation regarding Azure authentication for STT integrations. Sources: livekit-plugins/livekit-plugins-deepgram/livekit/plugins/deepgram/stt.py | Azure Speech STT integration guide | LiveKit Docs