Right now, changing the transcriber in a workflow applies globally. For multilingual workflows (English, Spanish, Arabic), we need different STT engines per node for best accuracy. Example: • English/Spanish → Nova-3 Medical • Arabic → Whisper Request: Let us set the transcriber per conversation node so each language or use case gets the most accurate, lowest-latency transcription — without having to create separate workflows. Why it matters: Boosts accuracy, reduces errors, and is critical for healthcare & multilingual environments.