Observability
Monitor the OM1 runtime with metrics, traces, and quality scoring.
OM1 exposes tooling to observe the runtime in real time — from low-level pipeline latencies to per-turn response quality.
This section covers:
Metrics Reference — the Prometheus metrics OM1 exports (ASR/LLM/TTS/VLM latencies, knowledge-base queries, response quality, and the internal HTTP proxy).
Tracer & Quality Scorer — record structured execution traces each turn and score response quality live.
Last updated
Was this helpful?