pipecat/changelog/4326.changed.md

• TranscriptionFrame.language and InterimTranscriptionFrame.language emitted by DeepgramFluxSTTService now reflect the language Deepgram detected for each turn (read from the languages field on Flux's TurnInfo event) instead of a statically configured value. On flux-general-multi this means per-turn accuracy for downstream consumers (e.g. TTS voice selection). On flux-general-en the field is absent in Flux responses, so emitted frames now carry language=None instead of the previously hardcoded Language.EN.