STT
Status: implemented.
Use STT when user audio must become text events for the pipeline.
Adapters implement stt.STT and emit speech events such as start of speech, interim transcript, final transcript, usage, and end of speech.
Configure app-level STT with:
export RTP_AGENT_STT_PROVIDER="deepgram"
export RTP_AGENT_STT_MODEL="nova-3"
export RTP_AGENT_STT_LANGUAGE="multi"
Provider-specific options vary widely. app.DefaultConfigFromEnv() exposes many RTP_AGENT_STT_* fields, but an option only matters when the selected adapter consumes it.
Evidence:
core/stt/stt.gocore/stt/stt_test.goadapter/*/stt.goapp/app.go
LinkedIn