I have a follow-up question as well: I want to measure and compare latency metrics (TTFB, processing time, etc.) across different STT/LLM/TTS models for performance benchmarking. Currently I'm using Langfuse + OpenTelemetry to collect these metrics: • STT: transcription delay, audio duration vs processing time • LLM: TTFT (Time to First Token), total completion time • TTS: TTFB, characters per second Is this the recommended approach? Or is there a better/easier/more accurate method you'd suggest? Thanks again!