Voice Agent observability with LangSmith 🌟
LangSmith now provides full observability for voice agents built on Gemini Live
“To take that agent to production safely, you need visibility into what your agent is doing.”
Google I/O demonstrated LangSmith integration with Gemini Live speech-to-speech agents, enabling transcript capture, tool call tracing, interruption event logging, and cost breakdowns for voice AI. This extends LangSmith's text-trace capabilities (evals, datasets, dashboards) to audio-native agents. It signals that voice agent observability is maturing into a production requirement as speech-to-speech models move beyond demos.